Merge branch 'main' into codex-split-attention-offload-guides

Update docs/design/feature/offloader/README.md
Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>
2026-08-15 07:00:22 +08:00 · 2026-08-13 23:37:08 +08:00 · 2026-08-13 23:37:01 +08:00 · 2026-08-12 11:35:07 +08:00 · 2026-08-12 09:30:34 +08:00 · 2026-08-12 08:55:41 +08:00
165 changed files with 1756 additions and 14086 deletions
@@ -76,9 +76,6 @@ steps:
          - tests/e2e/online_serving/test_qwen3_tts_customvoice.py
          - tests/e2e/offline_inference/test_qwen3_tts_customvoice.py
          - vllm_omni/model_executor/models/qwen3_tts/
-          - vllm_omni/model_executor/models/common/qwen3_code_predictor.py
-          - vllm_omni/model_executor/models/common/snake_activation.py
-          - vllm_omni/model_executor/models/common/alias_free_activation.py
          - vllm_omni/model_executor/stage_input_processors/qwen3_tts.py
          - vllm_omni/deploy/qwen3_tts.yaml
        commands:
@@ -95,9 +92,6 @@ steps:
          - tests/e2e/online_serving/test_qwen3_tts_base.py
          - tests/e2e/offline_inference/test_qwen3_tts_base.py
          - vllm_omni/model_executor/models/qwen3_tts/
-          - vllm_omni/model_executor/models/common/qwen3_code_predictor.py
-          - vllm_omni/model_executor/models/common/snake_activation.py
-          - vllm_omni/model_executor/models/common/alias_free_activation.py
          - vllm_omni/model_executor/stage_input_processors/qwen3_tts.py
          - vllm_omni/deploy/qwen3_tts.yaml
          - .buildkite/common/scripts/run_cov_split.sh
@@ -115,15 +109,27 @@ steps:
              -- -m 'advanced_model and cuda' --run-level 'advanced_model'
        mirror_hardwares: l4_1

+      - label: "TTS · Higgs-Audio-V3 Test"
+        source_file_dependencies:
+          - tests/e2e/online_serving/test_higgs_audio_v3.py
+          - vllm_omni/model_executor/models/higgs_audio_v3/
+          - vllm_omni/model_executor/stage_input_processors/higgs_audio_v3.py
+          - vllm_omni/deploy/higgs_multimodal_qwen3.yaml
+        commands:
+          - |
+            timeout 20m bash -c "
+              export VLLM_WORKER_MULTIPROC_METHOD=spawn
+              export VLLM_USE_DEEP_GEMM=0
+              export VLLM_MOE_USE_DEEP_GEMM=0
+              pytest -s -v tests/e2e/online_serving/test_higgs_audio_v3.py -m 'advanced_model' --run-level 'advanced_model'
+            "
+        mirror_hardwares: l4_1
+
      - label: "Omni · Qwen3-Omni Test"
        source_file_dependencies:
          - tests/e2e/offline_inference/test_qwen3_omni.py
          - tests/e2e/online_serving/test_qwen3_omni.py
          - vllm_omni/model_executor/models/qwen3_omni/
-          - vllm_omni/model_executor/models/common/qwen3_code_predictor.py
-          - vllm_omni/model_executor/models/common/snake_activation.py
-          - vllm_omni/model_executor/models/qwen2_5_omni/qwen2_5_omni_thinker.py
-          - vllm_omni/model_executor/models/qwen3_tts/cuda_graph_decoder_wrapper.py
          - vllm_omni/model_executor/stage_input_processors/qwen3_omni.py
          - vllm_omni/deploy/qwen3_omni_moe.yaml
        timeout_in_minutes: 50
@@ -135,11 +141,6 @@ steps:
        source_file_dependencies:
          - tests/e2e/offline_inference/test_minicpmo_4_5.py
          - tests/e2e/online_serving/test_minicpmo_4_5.py
-          - vllm_omni/model_executor/models/minicpmo_4_5/
-          - vllm_omni/model_executor/models/step_audio2/
-          - vllm_omni/model_executor/models/cosyvoice3/code2wav_core/
-          - vllm_omni/model_executor/models/cosyvoice3/utils.py
-          - vllm_omni/model_executor/stage_input_processors/minicpmo_4_5_omni.py
          - vllm_omni/deploy/minicpmo_4_5.yaml
        timeout_in_minutes: 50
        commands:
@@ -161,35 +162,70 @@ steps:
          - tests/e2e/online_serving/helpers/minicpmo_realtime_duplex_scenarios.py
          - tests/e2e/online_serving/run_minicpmo_realtime_duplex_multi_session.py
          - tests/e2e/online_serving/run_minicpmo_realtime_duplex_soft_interrupt.py
-          - vllm_omni/deploy/minicpmo_4_5.yaml
          - tests/e2e/online_serving/helpers/minicpmo_4_5_duplex.py
          - vllm_omni/deploy/minicpmo_4_5_duplex.yaml
          - vllm_omni/experimental/fullduplex/
          - vllm_omni/model_executor/models/minicpmo_4_5/
-          - vllm_omni/model_executor/models/step_audio2/
-          - vllm_omni/model_executor/models/cosyvoice3/code2wav_core/
-          - vllm_omni/model_executor/models/cosyvoice3/utils.py
          - vllm_omni/model_executor/stage_input_processors/minicpmo_4_5_omni.py
          - vllm_omni/core/sched/
          - vllm_omni/outputs/multimodal_accumulation.py
          - vllm_omni/engine/stage_init_utils.py
          - vllm_omni/engine/async_omni_engine.py
-          - vllm_omni/engine/orchestrator.py
          - vllm_omni/entrypoints/cli/serve.py
-          - vllm_omni/entrypoints/openai/api_server.py
-          - vllm_omni/entrypoints/openai/duplex_capability.py
-          - vllm_omni/config/stage_config.py
        timeout_in_minutes: 50
        commands:
          - pytest -s -v tests/e2e/online_serving/test_minicpmo_4_5_duplex.py -m 'advanced_model and cuda' --run-level 'advanced_model'
        mirror_hardwares: h100_1

+      - label: "TTS · VoxCPM2 Test"
+        source_file_dependencies:
+          - tests/e2e/offline_inference/test_voxcpm2_tts.py
+          - tests/e2e/online_serving/test_voxcpm2_tts.py
+          - vllm_omni/model_executor/models/voxcpm2/
+          - vllm_omni/deploy/voxcpm2.yaml
+        timeout_in_minutes: 30
+        commands:
+          - |
+            timeout 30m bash -c "
+              export VLLM_LOGGING_LEVEL=DEBUG
+              export VLLM_WORKER_MULTIPROC_METHOD=spawn
+              pytest -s -v tests/e2e/offline_inference/test_voxcpm2_tts.py tests/e2e/online_serving/test_voxcpm2_tts.py -m 'advanced_model and cuda' --run-level 'advanced_model'
+            "
+        mirror_hardwares: l4_1
+
+      - label: "Diffusion · Qwen Image Edit Test"
+        source_file_dependencies:
+          - tests/e2e/online_serving/test_qwen_image_edit.py
+          - vllm_omni/diffusion/models/qwen_image/
+        timeout_in_minutes: 20
+        commands:
+          - pytest -s -v tests/e2e/online_serving/test_qwen_image_edit.py -m 'advanced_model and cuda' --run-level 'advanced_model'
+        mirror_hardwares: h100_1
+      - label: "Diffusion · Bagel Test"
+        source_file_dependencies:
+          - tests/e2e/offline_inference/test_bagel.py
+          - tests/e2e/online_serving/test_bagel.py
+          - vllm_omni/model_executor/models/bagel/
+          - vllm_omni/diffusion/models/bagel/
+          - vllm_omni/model_executor/stage_input_processors/bagel.py
+          - vllm_omni/deploy/bagel.yaml
+          - .buildkite/common/scripts/run_cov_split.sh
+          - pyproject.toml
+        # 20 -> 40: the coverage split runs the model twice.
+        timeout_in_minutes: 40
+        commands:
+          - |
+            .buildkite/common/scripts/run_cov_split.sh \
+              --model-id bagel \
+              --offline tests/e2e/offline_inference/test_bagel.py \
+              --online tests/e2e/online_serving/test_bagel.py \
+              -- -m 'advanced_model and cuda' --run-level 'advanced_model'
+        mirror_hardwares: h100_1
+
      - label: "Diffusion · Qwen Image Test"
        source_file_dependencies:
          - tests/e2e/online_serving/test_qwen_image.py
          - vllm_omni/diffusion/models/qwen_image/
-          - vllm_omni/diffusion/models/dmd2/
-          - vllm_omni/diffusion/distributed/autoencoders/autoencoder_kl_qwenimage.py
        commands:
          - |
            timeout 40m bash -c '
@@ -198,12 +234,23 @@ steps:
            '
        mirror_hardwares: h100_1

+      - label: "Diffusion · Qwen Image Layered Test"
+        source_file_dependencies:
+          - tests/e2e/online_serving/test_qwen_image_layered.py
+          - vllm_omni/diffusion/models/qwen_image/
+        commands:
+          - |
+            timeout 40m bash -c "
+              export VLLM_IMAGE_FETCH_TIMEOUT=60
+              pytest -s -v tests/e2e/online_serving/test_qwen_image_layered.py -m 'advanced_model and cuda' --run-level 'advanced_model'
+            "
+        mirror_hardwares: h100_1
+
      - label: "Diffusion · Wan22 Test"
        source_file_dependencies:
          - tests/e2e/offline_inference/test_wan22_t2v.py
          - tests/e2e/online_serving/test_wan22_t2v.py
          - vllm_omni/diffusion/models/wan2_2/
-          - vllm_omni/diffusion/models/dmd2/
        commands:
          - |
            timeout 40m bash -c "
@@ -211,3 +258,21 @@ steps:
              pytest -s -v tests/e2e/offline_inference/test_wan22_t2v.py tests/e2e/online_serving/test_wan22_t2v.py -m 'advanced_model and cuda and diffusion' --run-level 'advanced_model'
            "
        mirror_hardwares: h100_1
+
+      - label: "Diffusion · Z Image Test"
+        source_file_dependencies:
+          - tests/e2e/offline_inference/test_z_image.py
+          - tests/e2e/online_serving/test_z_image.py
+          - vllm_omni/diffusion/models/z_image/
+        timeout_in_minutes: 30
+        commands:
+          - pytest -s -v tests/e2e/offline_inference/test_z_image.py tests/e2e/online_serving/test_z_image.py -m 'advanced_model and cuda and diffusion' --run-level 'advanced_model'
+        mirror_hardwares: l4_1
+      - label: "Diffusion · FLUX.2 Klein Test"
+        source_file_dependencies:
+          - tests/e2e/offline_inference/test_flux2_klein.py
+          - vllm_omni/diffusion/models/flux2_klein/
+        timeout_in_minutes: 30
+        commands:
+          - pytest -s -v tests/e2e/offline_inference/test_flux2_klein.py -m 'advanced_model and cuda and diffusion' --run-level 'advanced_model'
+        mirror_hardwares: l4_1
@@ -459,12 +459,6 @@ steps:
          - pytest -s -v tests/e2e/online_serving/test_wan_2_1_vace_expansion.py tests/e2e/online_serving/test_ltx.py -m "full_model and cuda" --run-level "full_model"
        mirror_hardwares: h100_2

-      - label: ":full_moon: Diffusion X2V · LTX-2.5 Online Test"
-        timeout_in_minutes: 60
-        commands:
-          - pytest -s -v tests/e2e/online_serving/test_ltx25.py -m "full_model and cuda" --run-level "full_model"
-        mirror_hardwares: h100_1
-
      - label: ":full_moon: Diffusion X2V · Wan2.2 Accuracy Test"
        timeout_in_minutes: 180
        commands:
@@ -493,13 +487,6 @@ steps:
          - pytest -s -v 'tests/e2e/accuracy/ltx/test_ltx_official_similarity.py::test_ltx_matches_official[ltx23_two_stage_layer_fused_i2v]' -m full_model --run-level full_model
        mirror_hardwares: h100_1

-      - label: ":full_moon: Diffusion X2V · LTX-2.5 Official Parity Test"
-        timeout_in_minutes: 180
-        commands:
-          - pytest -s -v tests/e2e/accuracy/ltx/test_ltx25_official_similarity.py::test_ltx25_distilled_two_stage_matches_official -m full_model --run-level full_model
-          - pytest -s -v tests/e2e/accuracy/ltx/test_ltx25_official_similarity.py::test_ltx25_full_matches_official -m full_model --run-level full_model
-        mirror_hardwares: h100_1
-
      - label: ":full_moon: Diffusion X2V · Perf Test"
        key: nightly-diffusion-x2v-performance
        timeout_in_minutes: 180
@@ -169,10 +169,6 @@ steps:
        source_file_dependencies:
          - tests/e2e/online_serving/test_qwen3_omni.py
          - vllm_omni/model_executor/models/qwen3_omni/
-          - vllm_omni/model_executor/models/common/qwen3_code_predictor.py
-          - vllm_omni/model_executor/models/common/snake_activation.py
-          - vllm_omni/model_executor/models/qwen2_5_omni/qwen2_5_omni_thinker.py
-          - vllm_omni/model_executor/models/qwen3_tts/cuda_graph_decoder_wrapper.py
          - vllm_omni/model_executor/stage_input_processors/qwen3_omni.py
          - vllm_omni/deploy/qwen3_omni_moe.yaml
        commands:
@@ -185,11 +181,6 @@ steps:
      - label: "Omni · MiniCPM-o 4.5 Offline Test"
        source_file_dependencies:
          - tests/e2e/offline_inference/test_minicpmo_4_5.py
-          - vllm_omni/model_executor/models/minicpmo_4_5/
-          - vllm_omni/model_executor/models/step_audio2/
-          - vllm_omni/model_executor/models/cosyvoice3/code2wav_core/
-          - vllm_omni/model_executor/models/cosyvoice3/utils.py
-          - vllm_omni/model_executor/stage_input_processors/minicpmo_4_5_omni.py
          - vllm_omni/deploy/minicpmo_4_5.yaml
        commands:
          - |
@@ -201,11 +192,6 @@ steps:
      - label: "Omni · MiniCPM-o 4.5 Online Test"
        source_file_dependencies:
          - tests/e2e/online_serving/test_minicpmo_4_5.py
-          - vllm_omni/model_executor/models/minicpmo_4_5/
-          - vllm_omni/model_executor/models/step_audio2/
-          - vllm_omni/model_executor/models/cosyvoice3/code2wav_core/
-          - vllm_omni/model_executor/models/cosyvoice3/utils.py
-          - vllm_omni/model_executor/stage_input_processors/minicpmo_4_5_omni.py
          - vllm_omni/deploy/minicpmo_4_5.yaml
        commands:
          - |
@@ -224,24 +210,16 @@ steps:
          - tests/e2e/online_serving/helpers/minicpmo_realtime_duplex_scenarios.py
          - tests/e2e/online_serving/run_minicpmo_realtime_duplex_multi_session.py
          - tests/e2e/online_serving/run_minicpmo_realtime_duplex_soft_interrupt.py
-          - vllm_omni/deploy/minicpmo_4_5.yaml
          - tests/e2e/online_serving/helpers/minicpmo_4_5_duplex.py
          - vllm_omni/deploy/minicpmo_4_5_duplex.yaml
          - vllm_omni/experimental/fullduplex/
          - vllm_omni/model_executor/models/minicpmo_4_5/
-          - vllm_omni/model_executor/models/step_audio2/
-          - vllm_omni/model_executor/models/cosyvoice3/code2wav_core/
-          - vllm_omni/model_executor/models/cosyvoice3/utils.py
          - vllm_omni/model_executor/stage_input_processors/minicpmo_4_5_omni.py
          - vllm_omni/core/sched/
          - vllm_omni/outputs/multimodal_accumulation.py
          - vllm_omni/engine/stage_init_utils.py
          - vllm_omni/engine/async_omni_engine.py
-          - vllm_omni/engine/orchestrator.py
          - vllm_omni/entrypoints/cli/serve.py
-          - vllm_omni/entrypoints/openai/api_server.py
-          - vllm_omni/entrypoints/openai/duplex_capability.py
-          - vllm_omni/config/stage_config.py
        commands:
          - |
            timeout 20m bash -c "
@@ -261,13 +239,25 @@ steps:
          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_minicpmo_4_5_ready.json -m 'core_model and cuda'
        mirror_hardwares: h100_1

+      - label: "TTS · VoxCPM2 Test"
+        source_file_dependencies:
+          - tests/e2e/online_serving/test_voxcpm2_tts.py
+          - vllm_omni/model_executor/models/voxcpm2/
+          - vllm_omni/deploy/voxcpm2.yaml
+        timeout_in_minutes: 30
+        commands:
+          - |
+            timeout 30m bash -c "
+              export VLLM_LOGGING_LEVEL=DEBUG
+              export VLLM_WORKER_MULTIPROC_METHOD=spawn
+              pytest -s -v tests/e2e/online_serving/test_voxcpm2_tts.py -m 'core_model' --run-level 'core_model'
+            "
+        mirror_hardwares: l4_1
+
      - label: "TTS · Qwen3-TTS CustomVoice Test"
        source_file_dependencies:
          - tests/e2e/online_serving/test_qwen3_tts_customvoice.py
          - vllm_omni/model_executor/models/qwen3_tts/
-          - vllm_omni/model_executor/models/common/qwen3_code_predictor.py
-          - vllm_omni/model_executor/models/common/snake_activation.py
-          - vllm_omni/model_executor/models/common/alias_free_activation.py
          - vllm_omni/model_executor/stage_input_processors/qwen3_tts.py
          - vllm_omni/deploy/qwen3_tts.yaml
        commands:
@@ -282,9 +272,6 @@ steps:
        source_file_dependencies:
          - tests/e2e/online_serving/test_qwen3_tts_base.py
          - vllm_omni/model_executor/models/qwen3_tts/
-          - vllm_omni/model_executor/models/common/qwen3_code_predictor.py
-          - vllm_omni/model_executor/models/common/snake_activation.py
-          - vllm_omni/model_executor/models/common/alias_free_activation.py
          - vllm_omni/model_executor/stage_input_processors/qwen3_tts.py
          - vllm_omni/deploy/qwen3_tts.yaml
        commands:
@@ -297,11 +284,52 @@ steps:
            "
        mirror_hardwares: l4_1

+      - label: "TTS · Higgs-Audio-V3 Test"
+        source_file_dependencies:
+          - tests/e2e/online_serving/test_higgs_audio_v3.py
+          - vllm_omni/model_executor/models/higgs_audio_v3/
+          - vllm_omni/model_executor/stage_input_processors/higgs_audio_v3.py
+          - vllm_omni/deploy/higgs_multimodal_qwen3.yaml
+        commands:
+          - |
+            timeout 20m bash -c "
+              export VLLM_WORKER_MULTIPROC_METHOD=spawn
+              export VLLM_USE_DEEP_GEMM=0
+              export VLLM_MOE_USE_DEEP_GEMM=0
+              pytest -s -v tests/e2e/online_serving/test_higgs_audio_v3.py -m 'core_model' --run-level 'core_model'
+            "
+        mirror_hardwares: l4_1
+
+      - label: "Diffusion · Qwen Image Edit Test"
+        source_file_dependencies:
+          - tests/e2e/online_serving/test_qwen_image_edit.py
+          - vllm_omni/diffusion/models/qwen_image/
+        commands:
+          - |
+            timeout 20m bash -c "
+              pytest -s -v tests/e2e/online_serving/test_qwen_image_edit.py -m 'core_model' --run-level 'core_model'
+            "
+        mirror_hardwares: h100_1
+
+      - label: "Diffusion · Bagel Test"
+        source_file_dependencies:
+          - tests/e2e/online_serving/test_bagel.py
+          - vllm_omni/model_executor/models/bagel/
+          - vllm_omni/diffusion/models/bagel/
+          - vllm_omni/model_executor/stage_input_processors/bagel.py
+          - vllm_omni/deploy/bagel.yaml
+        commands:
+          - |
+            timeout 40m bash -c "
+              export VLLM_IMAGE_FETCH_TIMEOUT=60
+              pytest -s -v tests/e2e/online_serving/test_bagel.py -m 'core_model' --run-level 'core_model'
+            "
+        mirror_hardwares: h100_1
+
      - label: "Diffusion · Wan22 Test"
        source_file_dependencies:
          - tests/e2e/online_serving/test_wan22_t2v.py
          - vllm_omni/diffusion/models/wan2_2/
-          - vllm_omni/diffusion/models/dmd2/
        commands:
          - |
            timeout 40m bash -c "
@@ -314,9 +342,6 @@ steps:
        source_file_dependencies:
          - tests/e2e/online_serving/test_cosmos3.py
          - vllm_omni/diffusion/models/cosmos3/
-          - vllm_omni/diffusion/distributed/autoencoders/autoencoder_kl_wan.py
-          - vllm_omni/experimental/world_models/
-          - vllm_omni/model_extras/cosmos3.py
        commands:
          - |
            timeout 40m bash -c "
@@ -329,8 +354,6 @@ steps:
        source_file_dependencies:
          - tests/e2e/online_serving/test_qwen_image.py
          - vllm_omni/diffusion/models/qwen_image/
-          - vllm_omni/diffusion/models/dmd2/
-          - vllm_omni/diffusion/distributed/autoencoders/autoencoder_kl_qwenimage.py
        commands:
          - |
            timeout 40m bash -c "
@@ -338,3 +361,15 @@ steps:
              pytest -s -v tests/e2e/online_serving/test_qwen_image.py -m 'core_model' --run-level 'core_model'
            "
        mirror_hardwares: h100_1
+
+      - label: "Diffusion · Qwen Image Layered Test"
+        source_file_dependencies:
+          - tests/e2e/online_serving/test_qwen_image_layered.py
+          - vllm_omni/diffusion/models/qwen_image/
+        commands:
+          - |
+            timeout 40m bash -c "
+              export VLLM_IMAGE_FETCH_TIMEOUT=60
+              pytest -s -v tests/e2e/online_serving/test_qwen_image_layered.py -m 'core_model' --run-level 'core_model'
+            "
+        mirror_hardwares: h100_1
@@ -108,22 +108,6 @@ steps:
            --run-level "full_model"
        mirror_hardwares: h100_1

-      # Own step: the image has no NanoCodec, so this file import-skips out of
-      # the mark-selected TTS steps above. The overrides keep nemo_toolkit[tts]
-      # from downgrading transformers; see
-      # examples/offline_inference/text_to_speech/README.md.
-      - label: "TTS · Gepard-1.0 · L4"
-        timeout_in_minutes: 60
-        commands:
-          - export VLLM_ALLOW_LONG_MAX_MODEL_LEN="1"
-          - printf '%s\n' "transformers>=5.5.3" "huggingface-hub>=1.0" > /tmp/gepard-overrides.txt
-          - uv pip install --python "$(python3 -c 'import sys; print(sys.executable)')" "nemo_toolkit[tts]==2.7.3" --overrides /tmp/gepard-overrides.txt
-          - >-
-            pytest -s -v tests/e2e/offline_inference/test_gepard_tts.py
-            -m "tts and cuda"
-            --run-level "full_model"
-        mirror_hardwares: l4_1
-
      - label: "Diffusion · L4"
        timeout_in_minutes: 60
        commands:
@@ -254,7 +254,6 @@ outputs/
 results/
 generated/
 output_*/
-/tests/e2e/accuracy/artifacts/

 # Configuration overrides
 configs/local.yaml
@@ -329,13 +329,11 @@ On PR builds, `upload_pipeline.py` logs `skip '…' (no changes under …)` for

 ### Per-model coverage

-Pilot (v1): one Merge-tier (L3) job in `.buildkite/cuda/test-merge.yml` uploads
-per-model, per-entry-mode coverage as Buildkite artifacts — "TTS · Qwen3-TTS Base
-Test". It runs on a single GPU so the pilot is cheap to reproduce. (A second
-diffusion pilot, "Diffusion · Bagel Test", was retired from merge; reintroduce
-another diffusion coverage job if a second pilot is needed again.) This is a
-pilot, not a rollout: no dashboard/visualization lives in this repo, and nothing
-is gated on coverage.
+Pilot (v1): two Merge-tier (L3) jobs in `.buildkite/cuda/test-merge.yml` upload
+per-model, per-entry-mode coverage as Buildkite artifacts — "Diffusion · Bagel
+Test" and "TTS · Qwen3-TTS Base Test". Both run on a single GPU so the pilot is
+cheap to reproduce. This is a pilot, not a rollout: no dashboard/visualization
+lives in this repo, and nothing is gated on coverage.

 **Naming convention**: `coverage-<model_id>-<mode>-<step_id>.xml.gz`, where
 `<model_id>` is the model's directory name under
@@ -412,8 +410,7 @@ runs a coverage job needs the same.
 List both `run_cov_split.sh` and `pyproject.toml` in every opted-in job's
 `source_file_dependencies` — both change what the job measures, so without them a
 change there is filtered out of normal PR builds and only surfaces in a later
-nightly. `tests/buildkite/test_upload_pipeline.py` covers the filter behavior with
-a synthetic job (it does not pin real merge labels).
+nightly. `tests/buildkite/test_upload_pipeline.py` asserts this for the pilots.
 Editing only the surrounding CI YAML still does not schedule them, so a PR that
 touches just the wiring needs a full E2E run (or the commands run on a GPU host)
 to produce artifacts. When checking a new model's
@@ -117,22 +117,7 @@ After (async D2H):

 ### Batch Split

-When `execute_model_batch` returns a `COMPUTE_DONE` with a single `async_output_id` for the entire batch, the executor splits it into per-request `async_output_id`s (format `{batch_id}/{request_id}`) via `_batch_split_map`. When `OUTPUT_READY` arrives, the pump extracts per-request results from the batch output and resolves each request's output future independently. If `OUTPUT_READY` arrives before `execute_batch` registers the split map, `execute_batch` directly adopts the early output.
-
-### Reliability, Lifecycle & Timeout Behavior
-
-1. **Drain Before Memory Release (`drain_async_outputs` / `_ASYNC_OUTPUT_DRAIN_TIMEOUT_S`)**:
-   When workers transition into sleep states (`handle_sleep_task`) or invoke memory-releasing methods, `WorkerProc.drain_async_outputs()` blocks up to `_ASYNC_OUTPUT_DRAIN_TIMEOUT_S` (10.0s) until all in-flight background D2H/SHM packing tasks finish. This acts as a synchronization barrier preventing worker sleep cycles from releasing device tensors while the side CUDA stream is still actively reading them. Draining typically completes within 10–50 ms under normal operation.
-
-2. **Timeout Diagnostics (`_ASYNC_OUTPUT_TIMEOUT`)**:
-   The engine waits for `wait_output_ready()` with a timeout of `_ASYNC_OUTPUT_TIMEOUT` (30.0s). If a timeout occurs, the executor logs detailed diagnostic bookkeeping (pending futures, cached outputs, and batch split maps) to identify whether the stall occurred in the worker, pump, or waiter.
-
-3. **Queue Serialization & Atomic Future Resolution**:
-   - `result_mq` writes are serialized with `_result_mq_lock` to protect the single-writer `MessageQueue` against concurrent writes from the worker main loop (`COMPUTE_DONE`) and background thread (`OUTPUT_READY`).
-   - `ResultPumpThread` unpacks shared-memory payloads first, and then atomically resolves waiting futures or populates `_completed_outputs` under `_futures_lock`, avoiding race conditions with concurrent `wait_output_ready()` invocations.
-
-4. **Storage-Aware IPC Packing**:
-   `pack_diffusion_output_shm()` evaluates `max(view_bytes, storage_bytes)` against `_SHM_TENSOR_THRESHOLD` (1 MB) to ensure that per-request slices sharing a large batch storage are moved to POSIX shared memory rather than serialized through pickle over the MessageQueue inline buffer.
+When `execute_model_batch` returns a `COMPUTE_DONE` with a single `async_output_id` for the entire batch, the executor splits it into per-request `async_output_id`s (format `{batch_id}/{request_id}`) via `_batch_split_map`. When `OUTPUT_READY` arrives, the pump extracts per-request results from the batch output and resolves each request's output future independently.

 ## Model Coverage

@@ -146,7 +131,6 @@ All models running in request-mode (`step_execution=False`, the default) automat
 |---|---|---|
 | **HunyuanImage-3.0** | Image | `False` |
 | **Qwen-Image** | Image | `True` |
-| **LTX-2.3** | Video / Audio | `False` |

 Other models with `step_execution=False` are also supported but not yet verified.

@@ -4,7 +4,7 @@ This document describes distributed layerwise offload (DLO) for diffusion
 models. DLO keeps only a small number of DiT blocks on the accelerator and
 streams the remaining blocks from host memory. The distributed backend can
 either shard those host-side weights across an existing parallel group or keep
-complete rank-local block sources and avoid an additional collective.
+the standard loader's rank-local weights and avoid an additional collective.

 For user-facing commands, see the
 [distributed layerwise offloading guide](../../../user_guide/diffusion/offloader/distributed_layerwise_offload.md)
@@ -14,20 +14,10 @@ and the [Cosmos3 recipe](https://github.com/vllm-project/vllm-omni/blob/main/rec

 DLO is implemented for multi-device diffusion execution. The default
 AllGather path is the primary path for DP and SP deployments. The
-`--dlo-no-use-allgather` path streams complete blocks independently and adds no
-DLO weight collective.
-
-Host storage is selected separately from the transfer protocol. The loader can
-produce a direct-checkpoint mmap plan for a proven-compatible runtime layout;
-otherwise it uses the ordinary loader. Consequently, no-AllGather replicas on
-the same node can share immutable checkpoint pages when direct mmap is
-selected, while the ordinary-loader fallback still keeps a private runtime
-copy per process.
-
-The Phase A shared-mmap support boundary is TP1. TP greater than one is an
-ordinary-loader compatibility path: DLO can consume the resulting TP-local
-tensors, but those configurations do not use checkpoint mmap and must not be
-used to claim shared-mmap host-memory savings.
+`--dlo-no-use-allgather` path is a rank-local compatibility mode: it is useful
+for standard-loader sharding, workstation bring-up, and systems where an
+additional DLO collective is undesirable, but it does not reduce host weight
+storage across ranks.

 The compatibility matrix below describes the current implementation. The
 unit-level guards are covered, but not every parallelism combination has a
@@ -51,36 +41,6 @@ TP is deliberately not used as DLO's AllGather group. HSDP has its own
 parameter-sharding lifecycle and is not allowed to be sharded a second time by
 DLO's AllGather path.

-### The loader owns host-weight planning
-
-Before it decides whether ordinary weight materialization can be skipped, the
-diffusion loader builds one `HostWeightPlan`. A direct-checkpoint mmap plan is
-accepted only when preflight proves all of the following:
-
- every required DiT parameter and persistent buffer has exactly one source;
- runtime names, checkpoint keys, shapes, and dtypes match;
- the runtime topology is TP1 without HSDP or online quantization; and
- every custom loader operation is represented by a loader-owned checkpoint
-  adapter.
-
-The exact plan object is handed to DLO. The backend does not rescan checkpoint
-files, repeat the capability decision, or reconstruct names from its block
-topology. If preflight fails, the loader materializes weights normally and DLO
-consumes those runtime tensors.
-
-The plan owns only dedicated DiT component sources. If a pipeline also exposes
-ordinary sources for a text encoder or another non-DiT component, the loader
-still consumes those sources and includes their loaded names in its strict
-coverage check. Only the source prefixes covered by the plan skip ordinary
-materialization. A source that mixes DiT and non-DiT weights fails closed to
-the complete ordinary-loader path because it cannot be skipped safely as a
-unit.
-
-This boundary keeps checkpoint semantics out of DLO and avoids model-pipeline
-flags such as `_supports_mmap_loading` or parameter attributes for mmap-only
-transforms. Model-specific direct-layout knowledge, when required, lives in a
-checkpoint adapter beside the ordinary loader.
-
 ### AllGather path

 With the default `dlo_use_allgather=True`, each rank stores approximately
@@ -101,15 +61,6 @@ Buffers:    [current slot]       [prefetch slot]       [current slot]
 The backend uses two shared device buffers, so accelerator weight residency is
 bounded by the largest streamed blocks rather than the complete model.

-When direct checkpoint mmap is selected, the checkpoint mappings are only the
-source used to prepare each rank's persistent shard. They can be closed after
-shard preparation. Across the AllGather group, those private shards total
-approximately one runtime model copy.
-
-An effective DLO group size of one performs no collective, even when
-`dlo_use_allgather=True`; it follows the rank-local transfer path described
-below.
-
 When DP is greater than one, the engine can process one request per DP rank in
 the same denoising wave. Because AllGather is a collective, all participating
 requests must take the same execution path at every denoising step.
@@ -117,20 +68,9 @@ requests must take the same execution path at every denoising step.
 ### Rank-local path without DLO AllGather

 With `--dlo-no-use-allgather`, DLO forces its internal offload shard size to
-one and streams complete blocks using H2D copies only. The host backing may be
-either a loader-approved checkpoint mapping or ordinary runtime tensors.
-
-For direct mmap, each process retains immutable safetensors views and uses two
-bounded pinned host staging slots. Processes on the same node that map the same
-files share physical checkpoint pages through the OS page cache. This removes
-the persistent private full-model copy per pure-DP process, but each process
-still packs and transfers every complete block. Sharing is node-local; each
-node has its own page cache.
-
-When direct mmap preflight fails, the regular model loader remains responsible
-for preparing each rank's weights, including TP-local tensors or HSDP-managed
-parameters. In that fallback, each pure-DP process keeps a private full runtime
-copy.
+one. The regular model loader remains responsible for preparing each rank's
+weights, including TP-local tensors or HSDP-managed parameters. DLO then
+streams those rank-local tensors block by block using H2D copies only.

 This mode means:

@@ -140,18 +80,21 @@ This mode means:
  not shard weights across SP ranks.
 - TP/HSDP/SP collectives, if configured, are not disabled by this flag; only
  DLO's additional weight AllGather is disabled.
- Pure DP deployments share one checkpoint-backed copy per node when direct
-  mmap is selected; the ordinary-loader fallback keeps one private runtime
-  copy per rank.
+- Pure DP deployments keep a full host-side model copy per rank, subject to
+  shared OS page-cache behavior.
 - The scheduler does not require a synchronized DP request wave for DLO.

+This path is intentionally not implemented by reusing the AllGather mmap
+loader. It relies on the standard loader so model-specific TP/HSDP transforms
+remain intact.
+
 ## Parallelism compatibility

 | Parallelism | DLO + AllGather | DLO without AllGather |
 |---|---|---|
-| **DP** | Supported primary path. DLO shards host weights across the DP group and can run DP multi-concurrency. | Supported rank-local path. Compatible TP1 replicas can share checkpoint pages on each node; fallback runtime tensors remain private. |
+| **DP** | Supported primary path. DLO shards host weights across the DP group and can run DP multi-concurrency. | Supported rank-local path. DP replicas remain independent; no DLO weight sharding or cross-DP DLO collective. |
 | **SP** | Supported in the implementation. With DP=1, DLO uses the SP group for host-weight sharding; SP still shards sequence/activation work. | SP remains active, but DLO keeps standard-loader rank-local weights and adds no SP weight collective. |
-| **TP > 1** | Outside the Phase A shared-mmap support scope. The loader falls back before mutation, preserves TP-local layouts, and DLO may apply DP/SP host sharding to those ordinary runtime tensors. | Outside the Phase A shared-mmap support scope. The ordinary TP-aware loader produces rank-local tensors, which DLO streams without an additional weight collective; DP replicas retain private runtime storage. |
+| **TP > 1** | Unsupported in the DLO mmap path. TP-aware loader callbacks are bypassed, so the backend rejects this configuration when it enters that path. | Standard loading is retained and TP-local tensors can be streamed. This is the intended compatibility path, but it still needs broader model and hardware validation. |
 | **HSDP** | Rejected. HSDP has already sharded parameters, so DLO AllGather would double-shard them. | Accepted by configuration. HSDP owns parameter sharding and its own gathers; DLO only stages rank-local parameters. End-to-end coverage is limited. |

 ### Combined dimensions
@@ -179,15 +122,11 @@ AllGather DP multi-concurrency requires:
 The no-AllGather path does not impose these DLO-specific synchronized-wave
 requirements.

-Direct checkpoint mmap can back either transfer path. It is currently limited
-to proven TP1, non-HSDP, non-online-quantized layouts. Other layouts use the
-ordinary loader. Online quantization remains incompatible with DLO AllGather;
-use `--dlo-no-use-allgather` or disable online quantization.
-
-A normalized runtime mmap cache, built through the ordinary loader, is the
-proposed general mechanism for sharing transformed TP or quantized layouts.
-That cache and its publication/lifecycle protocol are intentionally outside
-this phase; see [RFC #6195](https://github.com/vllm-project/vllm-omni/issues/6195).
+The mmap loader is used only by the supported DLO+AllGather path when the
+model has a compatible checkpoint layout. Online quantization is incompatible
+with that sharded mmap path; use `--dlo-no-use-allgather` or disable online
+quantization. Models that do not meet the mmap requirements use the regular
+loader path.

 ## Validation coverage

@@ -195,75 +134,15 @@ Current source-level validation includes:

 - HSDP + DLO + AllGather rejection;
 - HSDP + DLO without AllGather acceptance at configuration level;
- loader preflight fallback for TP, HSDP, online quantization, unknown custom
-  loaders, missing keys, and shape/dtype mismatches;
- exact loader-to-backend plan transfer and ordinary-loader fallback;
- rank-local mmap source retention, bounded two-slot staging, and adapter
-  transforms without parameter-side flags;
+- TP rejection in the DLO+AllGather mmap path;
 - resident-layer requests requiring no-AllGather;
 - DP request-wave validation for denoising-step compatibility;
 - sharding, double-buffer, AllGather-size, and heterogeneous-block regression
  tests.

-### B300 parallel-topology smoke matrix
-
-A four-GPU B300 smoke test covered MiniMax-H3 FL2VA with the same prompt, seed,
-CUDNN attention backend, 256x256 output, two denoising steps, and
-`dlo_resident_layers=0`. The TP2 rows used DiT DP2xTP2 with the text encoder and
-VAEs at TP1. They validate the ordinary-loader fallback only, not direct mmap
-or shared-mmap host-memory savings.
-
-| Configuration | Result | Warm E2E | Peak device memory | Host PSS |
-|---|---:|---:|---:|---:|
-| DP4xTP1 AllGather | Passed, 4 concurrent requests | 2.87 s / 4 requests | 13.84 GiB | 211.99 GiB |
-| DP4xTP1 no-AllGather | Passed, 1 request | 15.02 s | 13.23 GiB | 187.77 GiB |
-| DP2xTP2 AllGather | Passed, 2 concurrent requests | 4.16 s / 2 requests | 12.50 GiB | 211.97 GiB |
-| DP2xTP2 no-AllGather | Passed, 1 request | 3.51 s | 11.88 GiB | 314.01 GiB |
-
-Within each topology, the AllGather and no-AllGather video and audio outputs
-were byte-identical. All four runs completed without an `ERROR` or traceback
-and released their device allocations. For DP4xTP1, no-AllGather direct mmap
-reduced total PSS by 24.22 GiB (11.4%) and `Private_Dirty` from 211.33 to
-125.32 GiB (40.7%) relative to AllGather. For DP2xTP2, preflight selected the
-ordinary loader as designed; no-AllGather PSS was 314.01 GiB, about 48% above
-AllGather, because DP replicas did not share checkpoint-backed runtime
-weights. This is a functional and memory smoke test, not a production-quality
-performance or output-quality benchmark.
-
-### Host-memory measurement
-
-A two-worker MiniMax-H3 FL2VA measurement on one L20X node compared the
-ordinary-loader fallback with direct mmap. Both runs used
-DP=2, TP=1, no DLO AllGather, BF16 weights, two denoising steps, and a
-256x256 four-second request. The ordinary-loader workers were sampled after
-initialization. The mmap workers were sampled after one completed request, so
-the checkpoint working set had been faulted into the page cache; this is the
-more conservative point for mmap.
-
-The values below come from `/proc/<worker>/smaps_rollup` and include the whole
-worker, not only the DiT. The stable rank-to-rank difference comes from other
-pipeline components, so each worker should be compared with the same worker in
-the other storage mode.
-
-| Worker | Ordinary RSS | mmap RSS | Ordinary PSS | mmap PSS | PSS reduction |
-|---|---:|---:|---:|---:|---:|
-| DP worker 0 | 168.27 GiB | 132.76 GiB | 167.84 GiB | 101.43 GiB | 66.40 GiB |
-| DP worker 1 | 116.19 GiB | 79.97 GiB | 115.73 GiB | 48.64 GiB | 67.09 GiB |
-| **Two-worker total** | — | — | **283.56 GiB** | **150.08 GiB** | **133.48 GiB (47.1%)** |
-
-The direct-mmap workers each reported 62.45 GiB `Shared_Clean` but only
-31.20 GiB `Pss_File`, which is the proportional charge expected when the same
-resident checkpoint pages are mapped by two workers. `Private_Dirty` also fell
-from 167.53 to 70.24 GiB for worker 0 and from 115.40 to 17.44 GiB for worker
-1, a reduction of about 97–98 GiB per worker. RSS understates this benefit
-because it counts a shared physical page in every process that maps it; summed
-PSS is the appropriate node-memory comparison.
-
-The highest-value missing coverage is broader end-to-end numerical and
-lifecycle comparison against ordinary layerwise offload for DP+SP,
-HSDP+SP+no-AllGather, and TP greater than one across additional models and
-target CUDA/NCCL or CANN/HCCL hardware. That broader TP coverage does not
-change the Phase A direct-mmap TP1 support boundary.
+The highest-value missing coverage is end-to-end numerical comparison against
+ordinary layerwise offload for DP+SP, TP+no-AllGather, and HSDP+SP+no-AllGather
+on the target CUDA/NCCL or CANN/HCCL hardware.

 ## Recommendations

@@ -271,9 +150,7 @@ change the Phase A direct-mmap TP1 support boundary.
  scaling path.
 - Use **SP + DLO AllGather** for long-sequence workloads when DP concurrency is
  not the goal.
- Use **no-AllGather** when independent replica execution is required. TP1
-  direct-mmap deployments can share checkpoint pages per node; other layouts
-  retain the ordinary loader's private host memory behavior and are outside
-  the Phase A shared-mmap support scope.
+- Use **no-AllGather** to bring up TP or workstation PCIe configurations, with
+  the expectation of higher host-memory use and lower validation confidence.
 - Prefer **HSDP alone** for production HSDP deployments until the combined
  HSDP + DLO no-AllGather path has broader end-to-end coverage.
@@ -76,12 +76,6 @@ equivalent. `auto-round` and `auto_round` use the Omni INC/AutoRound adapter.
 ModelOpt checkpoint metadata is detected from its method, producer metadata, or
 `quant_algo`, then mapped to the corresponding vLLM ModelOpt configuration.

-Unless the input is a per-component dictionary, the resolved configuration is
-passed unchanged to every quantization-aware component constructed by the
-pipeline. It affects only layers supported by that component's quantization
-method; it does not rewrite arbitrary `torch.nn` modules. A per-component
-dictionary narrows this global scope through runtime layer-prefix routing.
-
 The factory uses Omni-specific overrides for methods whose diffusion behavior
 or platform dispatch is not provided by the vLLM registry. The current override
 set includes INT8, BitsAndBytes, MXFP8, MXFP4, MXFP4 dual-scale, and INC/
@@ -160,11 +154,9 @@ quantizable vLLM layers. Each method configuration is responsible for:
 Model code must keep precision-sensitive or unsupported modules unquantized.
 The shared `safe_quant_config()` helper is used by model integrations for
 normalization and modulation layers where a generic FP8 configuration would be
-unsafe. Encoders exposed through the quantization factory receive a global
-configuration just like the main transformer, but only their supported vLLM
-quantizable layers are affected. Ordinary `torch.nn` encoder layers, VAEs,
-schedulers, and tokenizers are not rewritten automatically. A component map can
-select a narrower scope, but it cannot add quantization support to a layer.
+unsafe. Text encoders, audio/vision encoders, VAEs, schedulers, and tokenizers
+are not implicitly quantized just because the main transformer is quantized;
+they require explicit component routing and method support.

 Weight quantization and quantized KV cache are separate paths. The former is
 resolved through `quantization_config`; runtime attention/KV-cache quantization
@@ -101,7 +101,7 @@ outputs = omni.generate(
 )

 # Access custom trajectory data
-output = outputs[0]
+output = outputs[0].request_output
 print(f"Trajectory timesteps shape: {output.metrics['trajectory_timesteps'].shape}")
 print(f"Trajectory latents shape: {output.latents.shape}")
 ```
@@ -47,7 +47,7 @@ if __name__ == "__main__":
    omni = Omni(model="Tongyi-MAI/Z-Image-Turbo")
    prompt = "a cup of coffee on the table"
    outputs = omni.generate(prompt)
-    images = outputs[0].images
+    images = outputs[0].request_output.images
    images[0].save("coffee.png")
 ```

@@ -75,7 +75,8 @@ if __name__ == "__main__":
    ]
    omni_outputs = omni.generate(prompts)
    for i_prompt, prompt_output in enumerate(omni_outputs):
-        this_images = prompt_output.images
+        this_request_output = prompt_output.request_output
+        this_images = this_request_output.images
        for i_image, image in enumerate(this_images):
            image.save(f"p{i_prompt}-img{i_image}.jpg")
            print("saved to", f"p{i_prompt}-img{i_image}.jpg")
@@ -43,12 +43,7 @@ th {
 | `Wan22VACEPipeline` | Wan2.2-VACE | `Pyros13/Wan2.2-VACE-Fun-A14B-Diffusers` | ✅︎ |   |   |   | — |
 | `LTX2Pipeline` | LTX-2 / LTX-2.3 one-stage T2V and I2V | `Lightricks/LTX-2`, `diffusers/LTX-2.3-Diffusers` | ✅︎ | ✅︎ | | | — |
 | `LTX2TwoStagePipeline` | LTX-2 / LTX-2.3 ordinary two-stage T2V and I2V | `Lightricks/LTX-2`, `diffusers/LTX-2.3-Diffusers` + matching Lightricks LoRA and upsampler | ✅︎ | ✅︎ | | | — |
-| `LTX2DistilledOneStagePipeline` | LTX-2 / LTX-2.3 merged-distilled one-stage T2V and I2V | `rootonchair/LTX-2-19b-distilled`, `diffusers/LTX-2.3-Distilled-Diffusers` | ✅︎ | ✅︎ | | | — |
-| `LTX2DistilledTwoStagePipeline` | LTX-2 / LTX-2.3 merged-distilled two-stage T2V and I2V | `rootonchair/LTX-2-19b-distilled`, `diffusers/LTX-2.3-Distilled-Diffusers` + matching Lightricks upsampler | ✅︎ | ✅︎ | | | — |
-| `LTX2Pipeline` | LTX-2.5 Full/SFT one-stage T2V and I2V | `Lightricks/LTX-2.5-Diffusers` | ✅︎ | | | | [Repository](https://github.com/vllm-project/vllm-omni/blob/main/recipes/LTX/LTX-2.5.md) |
-| `LTX2TwoStagePipeline` | LTX-2.5 Full/SFT two-stage T2V and I2V | `Lightricks/LTX-2.5-Diffusers` + `Lightricks/LTX-2.5` LoRA | ✅︎ | | | | [Repository](https://github.com/vllm-project/vllm-omni/blob/main/recipes/LTX/LTX-2.5.md) |
-| `LTX2DistilledOneStagePipeline` | LTX-2.5 merged-distilled one-stage T2V and I2V | `Lightricks/LTX-2.5-Diffusers` | ✅︎ | | | | [Repository](https://github.com/vllm-project/vllm-omni/blob/main/recipes/LTX/LTX-2.5.md) |
-| `LTX2DistilledTwoStagePipeline` | LTX-2.5 merged-distilled two-stage T2V and I2V | `Lightricks/LTX-2.5-Diffusers` | ✅︎ | | | | [Repository](https://github.com/vllm-project/vllm-omni/blob/main/recipes/LTX/LTX-2.5.md) |
+| `LTX2DistilledPipeline` | LTX-2 / LTX-2.3 full-distilled two-stage T2V and I2V | `rootonchair/LTX-2-19b-distilled`, `diffusers/LTX-2.3-Distilled-Diffusers` + `Lightricks/LTX-2.3` upsampler | ✅︎ | ✅︎ | | | — |
 | `LingBotVideoPipeline` | LingBot-Video dense and MoE T2I, T2V, TI2V | `robbyant/lingbot-video-dense-1.3b`, `robbyant/lingbot-video-moe-30b-a3b` | ✅︎ | | | | — |
 | `MiniMaxH3Pipeline` | MiniMax H3 T2VA, first/last-keyframe FL2VA, and mixed-reference Ref2VA | `MiniMaxAI/MiniMax-H3` | ✅︎ | ✅︎ | | | [Published](https://recipes.vllm.ai/MiniMaxAI/MiniMax-H3) |
 | `DreamZeroPipeline` | DreamZero-DROID | `GEAR-Dreams/DreamZero-DROID` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
@@ -103,7 +98,6 @@ th {
 |`DyninOmniForConditionalGeneration` | Dynin-Omni | `snu-aidas/Dynin-Omni` | ✅︎ | | | | — |
 | `MiniCPMO45OmniForConditionalGeneration` | MiniCPM-o 4.5 | `openbmb/MiniCPM-o-4_5` | ✅︎ | | ✅︎ | | — |
 | `ErnieImagePipeline` | ERNIE-Image | `baidu/ERNIE-Image`, `baidu/ERNIE-Image-Turbo` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
-| `GepardTalkerForConditionalGeneration` | Gepard-1.0 | `nineninesix/gepard-1.0` | ✅︎ | | | | — |
 |`HiDreamImagePipeline` | HiDream-I1-Full | `HiDream-ai/HiDream-I1-Full` | ✅︎ | ✅︎ | | | — |

 The `Recipe` column links the corresponding deployment recipe. `Published` is
@@ -3,9 +3,7 @@
 Distributed layerwise offloading (DLO) extends block streaming to multi-device
 deployments. With AllGather enabled, each rank stores roughly `1 / dp_size` of
 the host weights and reconstructs each layer at runtime. Without AllGather,
-each rank streams a complete block independently. Compatible TP1 deployments
-can share checkpoint-backed host pages among processes on the same node;
-otherwise DLO streams the ordinary loader's rank-local tensors.
+each rank streams its standard-loader rank-local weights independently.

 See the [DLO feature design](../../../design/feature/offloader/distributed_layerwise_offload.md)
 for the implementation contract and compatibility matrix.
@@ -61,48 +59,21 @@ omni = Omni(
 | `--enable-distributed-layerwise-offload` | Enable DLO | `false` |
 | `--data-parallel-size N` | DP ranks and AllGather weight-sharding group | `1` |
 | `--dlo-use-allgather` | Shard host weights and reconstruct with AllGather | `true` |
-| `--dlo-no-use-allgather` | Stream complete rank-local blocks without a DLO weight collective | `false` |
+| `--dlo-no-use-allgather` | Keep standard-loader rank-local weights | `false` |
 | `--dlo-resident-layers N` | Keep N leading main-DiT blocks on device; requires no-AllGather and model-declared resident paths | `0` |

-## Host-weight loading
+## mmap weight loading

-The diffusion loader chooses host storage before DLO is enabled. It first
-attempts to build a complete, validated direct-checkpoint mmap plan. If names,
-coverage, shape, dtype, topology, or loader-callback compatibility cannot be
-proven, it runs the ordinary model loader instead. DLO consumes that result and
-does not make a second checkpoint-compatibility decision.
-
-The shared-mmap optimization in this phase is supported only with TP1. TP
-greater than one falls back before model mutation to ordinary TP-aware loading.
-DLO may still consume those TP-local tensors, but this is a compatibility path:
-it does not share checkpoint-backed runtime weights across DP replicas and
-provides no shared-mmap host-memory guarantee.
-
-The mmap plan skips only dedicated DiT weight sources. Other component sources,
-such as a text encoder loaded through the shared diffusion loader, continue to
-use their ordinary component loader. A checkpoint source that mixes DiT and
-non-DiT weights falls back completely rather than leaving an unplanned
-component uninitialized.
-
-With direct checkpoint mmap, the loader:
+The DLO plus AllGather path:

 1. saves non-persistent buffers such as RoPE frequencies;
 2. moves the normally created transformer to the meta device;
 3. loads checkpoint tensors as mmap views backed by the shared OS page cache;
-4. applies any loader-owned bounded layout adapters while packing blocks;
-5. restores saved non-persistent buffers; and
-6. preserves `post_load_weights()` and `validate_loaded_weights()` lifecycle
-   hooks.
+4. calls model-specific `post_load_weights()` conversions; and
+5. restores the saved non-persistent buffers.

-For AllGather with a group larger than one, each process copies only its
-persistent shard and then releases the source mapping. For no-AllGather, each
-process keeps the mapping open and packs complete blocks through two bounded
-pinned staging slots. Processes mapping the same files on one node share the
-immutable pages; no-AllGather still performs a complete-block H2D copy in each
-process.
-
-When the effective DLO group size is one, `dlo_use_allgather=True` does not
-perform a collective and uses the same rank-local transfer behavior.
+This avoids one full checkpoint RSS copy per rank and does not require
+model-specific loading code.

 ## Declarative topology

@@ -132,22 +103,16 @@ must enter each collective.

 ## Limitations

- Direct checkpoint mmap currently requires TP1. TP greater than one is
-  outside the Phase A shared-mmap support scope and falls back before model
-  mutation to the ordinary TP-aware loader. DLO can stream that runtime layout,
-  but it provides no shared-mmap host-memory benefit or guarantee.
+- Online FP8 quantization is rejected with the DLO plus AllGather mmap path.
+  Use `--dlo-no-use-allgather` or disable online quantization.
+- Tensor parallel size greater than one is rejected in the mmap path because
+  it bypasses TP-aware loader callbacks. The no-AllGather path retains the
+  standard loader and remains experimental with TP.
 - HSDP plus AllGather is rejected to avoid double sharding. HSDP without
  AllGather has limited end-to-end validation.
- Online quantization uses the ordinary loader with no-AllGather. It remains
-  incompatible with DLO AllGather.
 - Resident leading layers require `--dlo-no-use-allgather` and a model
  `OffloadPlan` that declares eligible `resident_dit_paths`.
 - DP concurrency requires an explicit, identical inference-step count.

-Sharing transformed TP or quantized runtime layouts through a normalized mmap
-cache is a follow-up design in
-[RFC #6195](https://github.com/vllm-project/vllm-omni/issues/6195), not part of
-the direct-checkpoint path.
-
 See the [Cosmos3 DistOffload recipe](https://github.com/vllm-project/vllm-omni/blob/main/recipes/cosmos3/Cosmos3-DistOffload.md)
 for an end-to-end example.
@@ -161,7 +161,6 @@ The following tables show which models support each feature:
 | **Wan2.1-VACE**              |     ❌     |     ✅      |           ✅           |       ✅        |         ✅         |         ❌         |   ✅    |             ✅             |      ✅ (decode)      |       ❌        |        ❌         |
 | **LTX-2**                    |     ❌     |     ✅      |           ✅           |       ✅        |         ✅         |         ❌         |   ✅    |             ✅             |      ✅ (decode)      |       ❌        |        ❌         |
 | **LTX-2.3**                  |     ❌     |     ✅      |           ✅           |       ✅        |         ✅         |         ❌         |   ✅    |             ✅             |      ✅ (decode)      |       ❌        |        ❌         |
-| **LTX-2.5**                  |     ❌     | ❓ (one-stage) | ❓ (Ulysses only) | ✅ (Full only) |         ❓         |         ❌         |   ✅    |             ✅             |      ✅ (decode)      |    ❓ (FP8)     |        ❌         |
 | **Helios**                   |     ❌     |     ✅      |           ✅           |       ✅        |         ✅         |         ❌         |   ✅    |             ✅             |          ❌           |       ❌        |        ✅*        |
 | **HunyuanVideo-1.5 T2V I2V** |     ❌     |     ✅      |           ✅           |       ✅        |         ✅         |         ❌         |   ✅    |             ✅             |  ✅ (encode/decode)   |       ✅        |        ❌         |
 | **DreamID-Omni**             |     ❌     |     ❌      |           ❌           |       ✅        |         ❌         |         ❌         |   ✅    |             ✅             |          ❌           |       ❌        |        ❌         |
@@ -175,7 +174,7 @@ The following tables show which models support each feature:
 **Frame Interpolation Support**

 - **Supported**: Wan2.2 text-to-video, image-to-video, and TI2V pipelines
- **Not supported**: Wan2.1-VACE, LTX-2, LTX-2.3, LTX-2.5, Helios, HunyuanVideo-1.5, DreamID-Omni
+- **Not supported**: Wan2.1-VACE, LTX-2, LTX-2.3, Helios, HunyuanVideo-1.5, DreamID-Omni

 ### AudioGen

@@ -43,7 +43,7 @@ if __name__ == "__main__":
            ),
        ],
    )
-    outputs[0].images[0].save("output.png")
+    outputs[0].request_output.images[0].save("output.png")
 ```

 ## Image-to-Image (Image Editing)
@@ -72,7 +72,7 @@ if __name__ == "__main__":
            ),
        ],
    )
-    outputs[0].images[0].save("output.png")
+    outputs[0].request_output.images[0].save("output.png")
 ```

 ## Generation Parameters
@@ -55,7 +55,7 @@ if __name__ == "__main__":
    omni = Omni(model="Qwen/Qwen-Image")
    prompt = "a cup of coffee on the table"
    outputs = omni.generate(prompt)
-    images = outputs[0].images
+    images = outputs[0].request_output.images
    images[0].save("coffee.png")
 ```

@@ -179,7 +179,7 @@ if __name__ == "__main__":
    ]
    outputs = omni.generate(prompts)
    for i, output in enumerate(outputs):
-        output.images[0].save(f"{i}.jpg")
+        output.request_output.images[0].save(f"{i}.jpg")
 ```

 !!! info
@@ -207,7 +207,7 @@ if __name__ == "__main__":
        }
    ])
    for i, output in enumerate(outputs):
-        output.images[0].save(f"{i}.jpg")
+        output.request_output.images[0].save(f"{i}.jpg")
 ```

 You can also pass a negative prompt via the CLI argument `--negative-prompt`:
@@ -95,7 +95,7 @@ warmup_quack_fp8([(14040, 2048, 6144), (14040, 2048, 2048)])
 | HunyuanImage-3.0 | `tencent/HunyuanImage-3.0`, `tencent/HunyuanImage-3.0-Instruct` | Yes | Yes | All layers; use the Hunyuan stage config for multi-stage runs | None | |
 | HunyuanVideo-1.5 | `hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v`, `720p_t2v`, `480p_i2v` | Yes | Yes | All layers | None | |
 | Cosmos3 | `nvidia/Cosmos3-Nano`, `nvidia/Cosmos3-Super` | Yes | Not validated | All layers | None | |
-| MiniMax-H3 | `MiniMaxAI/MiniMax-H3` (`FL2VA` / `Ref2VA`) | Yes | Not validated | `quantization="fp8"` quantizes eligible DiT and text-encoder linears; mixed-precision input/output heads stay FP32 | None | ✅︎ |
+| MiniMax H3 | `MiniMaxAI/MiniMax-H3` | Yes | Not validated | DiT linears except patch, timestep, and final projections; not compatible with layerwise offload | None | |

 ### Multi-Stage Omni/TTS Model (Qwen3-Omni, Qwen3-TTS)

@@ -138,39 +138,6 @@ outputs = omni.generate(
 )
 ```

-### Global and per-component scope
-
-A plain, global quantization configuration is passed unchanged to every
-quantization-aware component constructed by a pipeline. This includes an
-eligible encoder when that encoder is implemented with vLLM quantizable
-layers; it does not rewrite arbitrary `torch.nn` modules. A structured
-component map is the only way to narrow that scope. Pipeline integrations that
-do not yet expose an encoder through the quantization factory remain DiT-only.
-
-For a pipeline that exposes both a transformer and a quantization-aware text
-encoder, the scope is:
-
-| Configuration | Transformer | Text encoder | Components without supported quantizable layers |
-|---------------|-------------|--------------|-------------------------------------------------|
-| `quantization="fp8"` | FP8 | FP8 | checkpoint precision |
-| `{"transformer": {"method": "fp8"}}` | FP8 | checkpoint precision | checkpoint precision |
-| `{"text_encoder": {"method": "fp8"}}` | checkpoint precision | FP8 | checkpoint precision |
-
-Use `quantization_config` for component-selective Python configuration. For
-example, quantize only the text decoder:
-
-```python
-omni = Omni(
-    model="<your-model>",
-    quantization_config={"text_encoder": {"method": "fp8"}},
-)
-```
-
-Component keys are runtime prefixes exposed by the pipeline integration; common
-keys include `transformer` and `text_encoder`. Entries may be combined, and
-`ignored_layers` can keep named eligible layers in checkpoint precision. Check
-the model recipe for supported components and their runtime prefixes.
-
 CLI:

 ```bash
@@ -96,7 +96,7 @@ if __name__ == "__main__":
        ),
    )
    from diffusers.utils import export_to_video
-    frames = outputs[0].images
+    frames = outputs[0].request_output.images
    export_to_video(frames, "quick_test_i2v.mp4", fps=16)
 ```

@@ -68,7 +68,6 @@ from vllm_omni.model_extras import (
 from vllm_omni.model_extras import (
    get_extra_body_params,
    get_model_class_name,
-    should_preserve_reference_image_size,
 )
 from vllm_omni.outputs import OmniRequestOutput
 from vllm_omni.platforms import current_omni_platform
@@ -455,13 +454,7 @@ def main():

    media_inputs: dict[str, Any] = {}
    if image is not None:
-        preserve_image_size = should_preserve_reference_image_size(
-            model_class_name,
-            model=args.model,
-        )
-        media_inputs["image"] = (
-            image if preserve_image_size else image.resize((width, height), PIL.Image.Resampling.LANCZOS)
-        )
+        media_inputs["image"] = image.resize((width, height), PIL.Image.Resampling.LANCZOS)
    if last_image is not None:
        media_inputs["last_image"] = last_image.resize((width, height), PIL.Image.Resampling.LANCZOS)
    if mask_image is not None:
@@ -335,7 +335,7 @@ def main(args):
    print(f"use_tts (tts_bos): {use_tts}")

    for stage_outputs in omni_generator:
-        output = stage_outputs
+        output = stage_outputs.request_output
        if stage_outputs.final_output_type == "text":
            request_id = output.request_id
            text_output = output.outputs[0].text
@@ -1,32 +1,21 @@
 # SPDX-License-Identifier: Apache-2.0
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""Validate MiniMax-H3 request-mode and DLO lifecycle behavior.
-
-The DLO modes accept any positive data-parallel size.  With AllGather enabled,
-the script submits one complete DP wave and verifies recovery after an invalid
-wave.  It also checks that every mode shuts down all worker processes cleanly.
+"""Reproduce MiniMax-H3 request-mode and DLO/DP2 lifecycle behavior.

 Examples:

    VLLM_WORKER_MULTIPROC_METHOD=spawn \
    VLLM_OMNI_VIDEO_SYNC_TIMEOUT=14400 \
    VLLM_OMNI_DLO_DP_WAVE_TIMEOUT=600 \
-    CUDA_VISIBLE_DEVICES=0,1,2,3 \
-    python examples/offline_inference/minimax_h3/dlo_lifecycle.py \
-        --model /path/to/MiniMax-H3/FL2VA --mode dlo --dp-size 4
-
-    # Rank-local mmap storage, with no DLO collective or DP wave scheduling.
-    VLLM_WORKER_MULTIPROC_METHOD=spawn \
-    VLLM_OMNI_VIDEO_SYNC_TIMEOUT=14400 \
    CUDA_VISIBLE_DEVICES=0,1 \
-    python examples/offline_inference/minimax_h3/dlo_lifecycle.py \
-        --model /path/to/MiniMax-H3/FL2VA --mode dlo-no-allgather --dp-size 2
+    python examples/offline_inference/minimax_h3/dlo_dp2_lifecycle.py \
+        --model /path/to/MiniMax-H3/FL2VA --mode dlo-dp2

    VLLM_WORKER_MULTIPROC_METHOD=spawn \
    VLLM_OMNI_VIDEO_SYNC_TIMEOUT=14400 \
    CUDA_VISIBLE_DEVICES=0,1 \
-    python examples/offline_inference/minimax_h3/dlo_lifecycle.py \
-        --model /path/to/MiniMax-H3/FL2VA --mode request --tp-size 2
+    python examples/offline_inference/minimax_h3/dlo_dp2_lifecycle.py \
+        --model /path/to/MiniMax-H3/FL2VA --mode request
 """

 from __future__ import annotations
@@ -51,33 +40,14 @@ DEFAULT_PROMPTS = (
 )


-def positive_int(value: str) -> int:
-    parsed = int(value)
-    if parsed < 1:
-        raise argparse.ArgumentTypeError("must be a positive integer")
-    return parsed
-
-
 def parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument("--model", required=True, help="Path to MiniMax-H3/FL2VA")
    parser.add_argument(
        "--mode",
-        choices=("dlo", "dlo-no-allgather", "request"),
+        choices=("dlo-dp2", "request"),
        required=True,
-        help="DLO (with or without AllGather) or ordinary non-DLO request mode",
-    )
-    parser.add_argument(
-        "--dp-size",
-        type=positive_int,
-        default=2,
-        help="DLO data-parallel size (for example 1, 2, 4, or 8; default: 2)",
-    )
-    parser.add_argument(
-        "--tp-size",
-        type=positive_int,
-        default=2,
-        help="Tensor-parallel size used only by request mode (default: 2)",
+        help="DLO AllGather DP2 or ordinary non-DLO TP2 request mode",
    )
    parser.add_argument("--steps", type=int, default=2)
    parser.add_argument("--duration", type=float, default=5.0)
@@ -90,11 +60,10 @@ def parse_args() -> argparse.Namespace:


 def engine_kwargs(args: argparse.Namespace) -> dict[str, Any]:
-    is_dlo = args.mode in {"dlo", "dlo-no-allgather"}
    common: dict[str, Any] = {
        "model": args.model,
        "trust_remote_code": True,
-        "num_gpus": args.dp_size if is_dlo else args.tp_size,
+        "num_gpus": 2,
        "usp": 1,
        "ring": 1,
        "vae_parallel_mode": "tile",
@@ -105,22 +74,22 @@ def engine_kwargs(args: argparse.Namespace) -> dict[str, Any]:
        "stage_init_timeout": args.init_timeout,
        "init_timeout": args.init_timeout,
    }
-    if is_dlo:
+    if args.mode == "dlo-dp2":
        common.update(
            tensor_parallel_size=1,
-            data_parallel_size=args.dp_size,
+            data_parallel_size=2,
            text_encoder_tp_size=1,
            vae_patch_parallel_size=1,
            enable_distributed_layerwise_offload=True,
-            dlo_use_allgather=args.mode == "dlo",
+            dlo_use_allgather=True,
            dlo_resident_layers=0,
        )
    else:
        common.update(
-            tensor_parallel_size=args.tp_size,
+            tensor_parallel_size=2,
            data_parallel_size=1,
-            text_encoder_tp_size=args.tp_size,
-            vae_patch_parallel_size=args.tp_size,
+            text_encoder_tp_size=2,
+            vae_patch_parallel_size=2,
            enable_distributed_layerwise_offload=False,
        )
    return common
@@ -203,8 +172,7 @@ async def run(args: argparse.Namespace) -> dict[str, Any]:
        "engine_kwargs": engine_kwargs(args),
    }
    try:
-        uses_dp_wave = args.mode == "dlo" and args.dp_size > 1
-        if uses_dp_wave:
+        if args.mode == "dlo-dp2":
            started = time.perf_counter()
            asymmetric = await asyncio.gather(
                generate_one(
@@ -214,15 +182,12 @@ async def run(args: argparse.Namespace) -> dict[str, Any]:
                    prompt="",
                    seed=1001,
                ),
-                *(
-                    generate_one(
-                        engine,
-                        args,
-                        request_id=f"invalid-peer-{index}",
-                        prompt=DEFAULT_PROMPTS[index % len(DEFAULT_PROMPTS)],
-                        seed=1001 + index,
-                    )
-                    for index in range(1, args.dp_size)
+                generate_one(
+                    engine,
+                    args,
+                    request_id="invalid-peer",
+                    prompt=DEFAULT_PROMPTS[0],
+                    seed=1002,
                ),
                return_exceptions=True,
            )
@@ -230,18 +195,18 @@ async def run(args: argparse.Namespace) -> dict[str, Any]:
            summary["asymmetric_errors"] = [
                f"{type(result).__name__}: {result}" for result in asymmetric if isinstance(result, BaseException)
            ]
-            if len(summary["asymmetric_errors"]) != args.dp_size:
-                raise RuntimeError("Every request in the asymmetric DP wave must fail before dispatch")
+            if len(summary["asymmetric_errors"]) != 2:
+                raise RuntimeError("Both requests in the asymmetric DP wave must fail before dispatch")

        started = time.perf_counter()
-        request_count = args.dp_size if uses_dp_wave else 1
+        request_count = 2 if args.mode == "dlo-dp2" else 1
        outputs = await asyncio.gather(
            *(
                generate_one(
                    engine,
                    args,
                    request_id=f"recovery-{index}",
-                    prompt=DEFAULT_PROMPTS[index % len(DEFAULT_PROMPTS)],
+                    prompt=DEFAULT_PROMPTS[index],
                    seed=2000 + index,
                )
                for index in range(request_count)
@@ -74,7 +74,7 @@ if __name__ == "__main__":
    omni = Omni(model="Qwen/Qwen-Image")
    prompt = "a cup of coffee on the table"
    outputs = omni.generate(prompt)
-    images = outputs[0].images
+    images = outputs[0].request_output.images
    images[0].save("coffee.png")
 ```

@@ -265,7 +265,7 @@ if __name__ == "__main__":
    ]
    outputs = omni.generate(prompts)
    for i, output in enumerate(outputs):
-        output.images[0].save(f"{i}.jpg")
+        output.request_output.images[0].save(f"{i}.jpg")
 ```

 !!! info
@@ -301,7 +301,7 @@ if __name__ == "__main__":
        }
    ])
    for i, output in enumerate(outputs):
-        output.images[0].save(f"{i}.jpg")
+        output.request_output.images[0].save(f"{i}.jpg")
 ```

 You can also pass a negative prompt via the CLI argument `--negative-prompt`:
@@ -16,7 +16,6 @@ list of supported architectures across all modalities, see
 |---|---|---|---|---|---|---|
 | CosyVoice3 | `FunAudioLLM/Fun-CosyVoice3-0.5B-2512` | 2 (talker + code2wav) | ✓ | ✓ | — | 24 kHz |
 | Fish Speech S2 Pro | `fishaudio/s2-pro` | dual-AR | ✓ | ✓ | — | 44.1 kHz |
-| Gepard-1.0 | `nineninesix/gepard-1.0` | single (native AR + NanoCodec) | — (zero-shot; cloning WIP) | — (serving WIP) | zero-shot | 22.05 kHz |
 | GLM-TTS | `zai-org/GLM-TTS` | 2 (AR + DiT) | ✓ (required) | ✓ | — | 24 kHz |
 | Ming-omni-tts | `inclusionAI/Ming-omni-tts-0.5B` | 2 (AR + audio VAE) | ✓ | ✓ | style / IP / dialect / TTA / podcast | 44.1 kHz |
 | Ming-flash-omni-TTS | `Jonathan1909/Ming-flash-omni-2.0` | single (talker only) | — (caption-controlled) | — | style / IP / basic captions | 44.1 kHz |
@@ -396,58 +395,6 @@ python examples/offline_inference/text_to_speech/qwen3_tts/end2end.py \

 ---

-## Gepard-1.0
-
-Single-stage native AR TTS at 22.05 kHz. Pipeline: `Qwen3.5 backbone → 32 FSQ codebook heads (one frame/step) → NeMo NanoCodec`. Backbone runs under vLLM paged attention; the 32-head sampling + learned embedding feedback ride the native-AR runner hooks. `enforce_eager` (CUDA graph is a perf follow-up).
-
-### Prerequisites
-The NanoCodec decoder needs NeMo (NVIDIA Open Model License), installed separately:
-
-```bash
-uv pip install "nemo_toolkit[tts]==2.7.3" \
-    --overrides <(printf '%s\n' "transformers>=5.5.3" "huggingface-hub>=1.0")
-```
-
-The overrides are what make that command work. `nemo_toolkit[tts]` declares
-`transformers~=4.57.0`, so a plain install downgrades the version
-`requirements/common.txt` pins and takes `huggingface-hub` back to 0.x with it — after
-which `import vllm` fails. Both packages are named here rather than left to resolve on
-their own: transformers 5.x happens to require `huggingface-hub` 1.x, but relying on
-that would make the second repair an accident of the first. `--constraint` cannot be
-used instead, since it narrows a version range rather than overriding a dependency a
-package declares, and the two sets never intersect. NeMo 2.7.3 decodes correctly
-against transformers 5.x, so nothing is lost by holding it there.
-
-(The same clash is why this is documented here rather than declared in
-`pyproject.toml`: an extra resolves together with the base dependencies, so
-`vllm-omni[gepard]` would be unsatisfiable rather than opt-in. SoulX-Singer English
-SVS documents `nemo_toolkit[asr]` the same way.)
-
-On a host whose CUDA toolkit cannot build kernels for the GPU — no `nvcc`/`ninja`, or a
-consumer Blackwell (`sm_120`) card — also set `VLLM_USE_FLASHINFER_SAMPLER=0`. vLLM's
-sampler JIT-compiles through FlashInfer on first use, and where that build fails the engine
-dies inside `profile_run` long after every import has succeeded, which reads as a model
-failure rather than an environment one. The PyTorch sampler it falls back to costs nothing
-here.
-
-### Quick start (zero-shot, default voice)
-```bash
-python examples/offline_inference/text_to_speech/gepard/end2end.py \
-    --text "Hello, this is a Gepard demo."
-```
-
-### Voice cloning
-Not yet — PR1 is zero-shot only (the learned `null_prefix` default voice). Reference-audio cloning (`ref_compressor` speaker prefix) is a follow-up PR.
-
-### Notes
- Output: 22.05 kHz mono WAV.
- First run downloads two checkpoints, not one: the model itself, and the NanoCodec decoder it names in `codec_id` (`nvidia/nemo-nano-codec-22khz-1.89kbps-21.5fps` by default). Both are fetched automatically; expect the first startup to be correspondingly slower.
- Offline only for now; the `/v1/audio/speech` adapter is a follow-up PR.
- Deploy config: `vllm_omni/deploy/gepard.yaml` (the example's default; copy it and pass `--deploy-config` to change it).
- Generation length and reproducibility are stage settings, not script flags. One output token is one audio frame, so `max_tokens` in the YAML is the frame budget; `seed` makes the in-model 32-head sampling reproducible. The script deliberately passes no `SamplingParams`: one supplied by a caller replaces the stage defaults wholesale rather than merging, which would drop the pipeline's stop token and run every request to `max_tokens`.
- Text length: short texts are repeated internally to match the training layout (the checkpoint's `text_repetition` block); the upper bound is the stage's `max_model_len`, enforced by the engine. Empty text is rejected rather than voiced.
- `VLLM_GEPARD_GREEDY=1` swaps the 32-head Gumbel-max sampling for argmax, for reproducible comparisons that do not depend on a seed.
-
 ## VoxCPM2

 Single-stage native AR TTS at 48 kHz. Pipeline: `feat_encoder → MiniCPM4 → FSQ → residual_lm → LocDiT → AudioVAE`.
@@ -1,139 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""Offline inference example for Gepard-1.0 TTS via vLLM-Omni.
-
-Single-stage native-AR pipeline: a Qwen3.5 backbone samples one 32-code FSQ
-frame per step; the NeMo NanoCodec decodes the frames to a 22.05 kHz mono
-waveform. Zero-shot: the default learned voice, no reference audio.
-
-Generation length and seed are stage settings rather than flags here. One
-output token is one audio frame, so ``max_tokens`` in the deploy YAML is the
-frame budget; pass a copy of the YAML via ``--deploy-config`` to change it.
-
-Usage:
-  python end2end.py --text "Hello, this is Gepard speaking."
-"""
-
-from __future__ import annotations
-
-import os
-from pathlib import Path
-
-import soundfile as sf
-import torch
-
-import vllm_omni
-from vllm_omni.utils.tracking_parser import TrackingArgumentParser
-
-os.environ.setdefault("VLLM_WORKER_MULTIPROC_METHOD", "spawn")
-
-from vllm_omni import Omni  # noqa: E402
-
-MODEL = "nineninesix/gepard-1.0"
-SAMPLE_RATE = 22050
-# Absolute: the loader hands the path straight to OmegaConf, which resolves a
-# bare name against the current working directory, not the package.
-DEFAULT_DEPLOY_CONFIG = str(Path(vllm_omni.__file__).resolve().parent / "deploy" / "gepard.yaml")
-
-
-def build_request(text: str) -> dict:
-    """Build an Omni request payload for Gepard (zero-shot, no ref audio).
-
-    The model only consumes the [speaker slots, SOT, text, EOT, SOS] layout, so
-    a bare text prompt would have no speaker slots and no SOS.
-    """
-    from transformers import AutoTokenizer
-
-    from vllm_omni.model_executor.models.gepard.configuration_gepard import GepardConfig
-    from vllm_omni.model_executor.models.gepard.prompt import build_gepard_prompt_ids
-
-    # From the checkpoint's own sidecar, which is what the worker builds its
-    # config from too. A layout that disagrees with the trained one does not
-    # fail, it just degrades the speech.
-    cfg = GepardConfig.from_checkpoint(MODEL)
-    tokenizer = AutoTokenizer.from_pretrained(MODEL, trust_remote_code=True)
-    prompt_token_ids = build_gepard_prompt_ids(
-        tokenizer(text, add_special_tokens=False)["input_ids"],
-        config=cfg,
-    )
-    return {
-        "prompt_token_ids": prompt_token_ids,
-        "additional_information": {"text": [text]},
-    }
-
-
-def save_audio(waveform: torch.Tensor, path: str, sample_rate: int = SAMPLE_RATE) -> None:
-    sf.write(path, waveform.float().numpy(), sample_rate)
-    seconds = waveform.numel() / sample_rate
-    print(
-        f"  Saved {path} ({tuple(waveform.shape)}, {sample_rate} Hz, "
-        f"{seconds:.2f}s, peak {float(waveform.abs().max()):.3f})"
-    )
-
-
-def main(args) -> None:
-    omni = Omni(
-        model=MODEL,
-        deploy_config=args.deploy_config,
-        stage_init_timeout=args.stage_init_timeout,
-        init_timeout=args.init_timeout,
-    )
-
-    # No explicit SamplingParams: one supplied here would replace the stage
-    # defaults rather than merge over them, dropping the pipeline's stop token.
-    output_dir = Path(args.output_dir)
-    output_dir.mkdir(parents=True, exist_ok=True)
-
-    print(f"Synthesizing: {args.text!r}")
-    inputs = build_request(text=args.text)
-
-    for i, stage_outputs in enumerate(omni.generate(inputs)):
-        for j, out in enumerate(stage_outputs.outputs):
-            mm = out.multimodal_output
-            if mm is None:
-                print(f"  [req {i}] No audio output.")
-                continue
-            # The consolidation path renames "model_outputs" to "audio".
-            audio = mm.get("audio")
-            if audio is None:
-                audio = mm.get("model_outputs")
-            if isinstance(audio, list):
-                audio = audio[0] if len(audio) else None
-            if audio is None:
-                print(f"  [req {i}] No waveform in multimodal_output (keys: {list(mm)}).")
-                continue
-            sr_tensor = mm.get("sr")
-            sr = int(sr_tensor.item()) if hasattr(sr_tensor, "item") else SAMPLE_RATE
-            save_audio(audio.cpu(), str(output_dir / f"output_{i}_{j}.wav"), sr)
-
-    print("Done.")
-
-
-def parse_args():
-    parser = TrackingArgumentParser(description="Gepard-1.0 offline TTS inference (zero-shot)")
-    parser.add_argument("--text", default="Hello, this is Gepard speaking.", help="Text to synthesize.")
-    parser.add_argument(
-        "--output-dir",
-        default=os.path.join(
-            os.environ.get("XDG_CACHE_HOME", os.path.join(os.path.expanduser("~"), ".cache")),
-            "gepard_output",
-        ),
-        help="Directory for WAV outputs (default: ~/.cache/gepard_output).",
-    )
-    parser.add_argument(
-        "--deploy-config",
-        default=DEFAULT_DEPLOY_CONFIG,
-        help=(
-            "Path to the deploy YAML (default: the packaged vllm_omni/deploy/gepard.yaml). This must "
-            "stay set: without the YAML's `pipeline: gepard` pin the architectures fallback routes "
-            "Qwen3_5ForCausalLM to the diffusion registry. Copy it to change max_tokens or seed."
-        ),
-    )
-    # The upstream defaults are too tight for a cold start.
-    parser.add_argument("--stage-init-timeout", type=int, default=900)
-    parser.add_argument("--init-timeout", type=int, default=1800)
-    return parser.parse_args()
-
-
-if __name__ == "__main__":
-    main(parse_args())
@@ -570,8 +570,6 @@ def main():
        num_frames=args.num_frames,
        extra_args=extra_args,
    )
-    if args.frame_rate is not None:
-        sampling_kwargs["frame_rate"] = args.frame_rate
    if args.guidance_scale_high is not None:
        sampling_kwargs["guidance_scale_2"] = args.guidance_scale_high

@@ -1,175 +0,0 @@
-# LTX-2.5
-
-> Text-to-video and first-frame image-to-video generation with synchronized audio
-
-## Overview
-
-vLLM-Omni supports the gated
-[`Lightricks/LTX-2.5-Diffusers`](https://huggingface.co/Lightricks/LTX-2.5-Diffusers)
-checkpoint for T2V and first-frame I2V. Every output includes synchronized
-48 kHz stereo audio.
-
-The raw [`Lightricks/LTX-2.5`](https://huggingface.co/Lightricks/LTX-2.5)
-repository is not directly loadable with `--model`. It supplies the official
-upsampler and LoRA sidecars used by the Full/SFT two-stage pipeline, so accept
-both model licenses and authenticate before first use.
-
-## Choose a pipeline
-
-| Pipeline | Mode | Default size | Schedule |
-|---|---|---:|---:|
-| `LTX2Pipeline` | Full/SFT one-stage | 960x544 | 30 steps |
-| `LTX2TwoStagePipeline` | Full/SFT two-stage | 1920x1088 | 30 + 3 steps |
-| `LTX2DistilledOneStagePipeline` | Distilled one-stage | 960x544 | 8 steps |
-| `LTX2DistilledTwoStagePipeline` | Distilled two-stage | 1920x1088 | 8 + 3 steps |
-
-Both two-stage pipelines generate at half resolution, apply the official x2
-latent upsampler, and run a three-step refinement stage. Select the class with
-`--model-class-name`; no `--task-type` flag is required. Supplying one initial
-image selects I2V, while omitting it selects T2V.
-
-## Prerequisites
-
-```bash
-hf auth login
-export MODEL=Lightricks/LTX-2.5-Diffusers
-```
-
-Install matching vLLM and vLLM-Omni versions, and ensure `ffmpeg` and
-`ffprobe` are on `PATH`. I2V requires PyAV backed by an FFmpeg build with the
-`libx264` encoder.
-
-## Hardware
-
-| GPU | Status | Recommended scope |
-|---|---|---|
-| NVIDIA B300 | Verified | All four canonical pipelines |
-| NVIDIA B200 or H200 | Capacity-based recommendation; not yet verified | All four canonical pipelines |
-| NVIDIA GB200 or GB300 | Capacity-based recommendation; not yet verified | All four canonical pipelines |
-| NVIDIA H100 80 GB | FP8 recipe | Distilled one-stage at 960x544 |
-
-The 1920x1088 two-stage examples require about 114 GB of peak GPU memory.
-80 GB GPUs do not have enough safety margin for the canonical two-stage
-configuration; use a larger GPU or reduce the output configuration.
-
-### H100 80 GB FP8
-
-Use the distilled one-stage pipeline with FP8 and cuDNN attention:
-
-```bash
-vllm serve Lightricks/LTX-2.5-Diffusers \
-  --omni \
-  --model-class-name LTX2DistilledOneStagePipeline \
-  --quantization fp8 \
-  --diffusion-attention-backend CUDNN_ATTN \
-  --host 0.0.0.0 \
-  --port 8000
-```
-
-Use the T2V or I2V request below after the server is ready.
-
-## Offline inference
-
-Choose values from the pipeline table. For example, the distilled two-stage
-path uses:
-
-```bash
-export PIPELINE=LTX2DistilledTwoStagePipeline
-export WIDTH=1920
-export HEIGHT=1088
-export STEPS=8
-```
-
-### Text-to-video
-
-```bash
-python examples/offline_inference/text_to_video/text_to_video.py \
-  --model "${MODEL}" \
-  --model-class-name "${PIPELINE}" \
-  --prompt "A cinematic shot of a red fox walking through a snowy forest at dawn, the camera tracking alongside, snow crunching underfoot." \
-  --width "${WIDTH}" \
-  --height "${HEIGHT}" \
-  --num-frames 121 \
-  --num-inference-steps "${STEPS}" \
-  --frame-rate 24 \
-  --fps 24 \
-  --seed 42 \
-  --output ltx25-t2v.mp4
-```
-
-### First-frame image-to-video
-
-```bash
-python examples/offline_inference/image_to_video/image_to_video.py \
-  --model "${MODEL}" \
-  --model-class-name "${PIPELINE}" \
-  --image /absolute/path/to/first-frame.png \
-  --prompt "The red fox walks forward while the camera tracks alongside." \
-  --width "${WIDTH}" \
-  --height "${HEIGHT}" \
-  --num-frames 121 \
-  --num-inference-steps "${STEPS}" \
-  --frame-rate 24 \
-  --fps 24 \
-  --seed 42 \
-  --output ltx25-i2v.mp4
-```
-
-LTX-2.5 uses the official CRF-18 first-frame conditioning path by default.
-
-## Online serving
-
-Start one server for the selected pipeline:
-
-```bash
-vllm serve "${MODEL}" \
-  --omni \
-  --model-class-name "${PIPELINE}" \
-  --host 0.0.0.0 \
-  --port 8000 \
-  --stage-init-timeout 900
-```
-
-T2V request:
-
-```bash
-curl -sS --fail-with-body \
-  -X POST http://127.0.0.1:8000/v1/videos/sync \
-  -F 'prompt=A cinematic shot of a red fox walking through a snowy forest at dawn, the camera tracking alongside, snow crunching underfoot.' \
-  -o ltx25-online-t2v.mp4
-```
-
-For I2V, add exactly one first frame:
-
-```bash
-export FIRST_FRAME=/absolute/path/to/first-frame.png
-
-curl -sS --fail-with-body \
-  -X POST http://127.0.0.1:8000/v1/videos/sync \
-  -F "input_reference=@${FIRST_FRAME};type=image/png" \
-  -F 'prompt=The red fox walks forward while the camera tracks alongside.' \
-  -o ltx25-online-i2v.mp4
-```
-
-Restart the server after changing `PIPELINE` because each class selects a
-different weight layout and execution topology.
-
-## Constraints
-
- Online `width`, `height`, `num_frames`, `fps`, and `num_inference_steps`
-  are optional and use the selected pipeline defaults when omitted. `seed` is
-  optional and random when omitted.
- `num_frames` must be `8k+1` when overridden.
- One-stage width and height must be divisible by 32; two-stage final
-  dimensions must be divisible by 64.
- Full/SFT pipelines accept negative prompts. Distilled pipelines are
-  positive-only and reject negative prompts.
- I2V accepts exactly one initial image per prompt.
- The supported `--model` value is `Lightricks/LTX-2.5-Diffusers`.
-
-## References
-
- [LTX-2.5-Diffusers model card](https://huggingface.co/Lightricks/LTX-2.5-Diffusers)
- [Official LTX-2 repository](https://github.com/Lightricks/LTX-2)
- [Generic T2V example](../../examples/offline_inference/text_to_video/text_to_video.py)
- [Generic I2V example](../../examples/offline_inference/image_to_video/image_to_video.py)
@@ -8,12 +8,10 @@
 |---|---|---|
 | `LTX2Pipeline` | LTX-2 one-stage T2V/I2V | `Lightricks/LTX-2` |
 | `LTX2TwoStagePipeline` | LTX-2 ordinary two-stage T2V/I2V | `Lightricks/LTX-2` |
-| `LTX2DistilledOneStagePipeline` | LTX-2 merged-distilled one-stage T2V/I2V | `rootonchair/LTX-2-19b-distilled` |
-| `LTX2DistilledTwoStagePipeline` | LTX-2 merged-distilled two-stage T2V/I2V | `rootonchair/LTX-2-19b-distilled` |
+| `LTX2DistilledPipeline` | LTX-2 full-distilled two-stage T2V/I2V | `rootonchair/LTX-2-19b-distilled` |
 | `LTX2Pipeline` | LTX-2.3 one-stage T2V/I2V | `diffusers/LTX-2.3-Diffusers` |
 | `LTX2TwoStagePipeline` | LTX-2.3 ordinary two-stage T2V/I2V | `diffusers/LTX-2.3-Diffusers`<br>`Lightricks/LTX-2.3` |
-| `LTX2DistilledOneStagePipeline` | LTX-2.3 merged-distilled one-stage T2V/I2V | `diffusers/LTX-2.3-Distilled-Diffusers` |
-| `LTX2DistilledTwoStagePipeline` | LTX-2.3 merged-distilled two-stage T2V/I2V | `diffusers/LTX-2.3-Distilled-Diffusers`<br>`Lightricks/LTX-2.3` |
+| `LTX2DistilledPipeline` | LTX-2.3 full-distilled two-stage T2V/I2V | `diffusers/LTX-2.3-Distilled-Diffusers`<br>`Lightricks/LTX-2.3` |

 Repositories in the table are download units. A full pipeline repository
 contains the Transformer, text encoder, connectors, VAEs, vocoder, scheduler,
@@ -28,12 +26,9 @@ image selects I2V. Both one-stage repositories declare this class, so
 the raw `Lightricks/LTX-2.3` safetensors repository is not directly loadable.

 `LTX2TwoStagePipeline` samples the regular model at half resolution, upsamples,
-then refines with the distilled LoRA. `LTX2DistilledOneStagePipeline` uses a
-merged distilled Transformer without upsampling, while
-`LTX2DistilledTwoStagePipeline` uses it in both stages. All entries support
-T2V and I2V; select their class explicitly. The deprecated
-`LTX2DistilledPipeline` name remains an alias for
-`LTX2DistilledTwoStagePipeline`.
+then refines with the distilled LoRA. `LTX2DistilledPipeline` uses a fully
+merged distilled Transformer in both stages. Both entries support T2V and I2V;
+select their class explicitly.

 ## API Migration

@@ -57,8 +52,8 @@ The consolidation also removes these registry names without aliases:
 | `LTX23Pipeline` | `LTX2Pipeline`; checkpoint metadata selects LTX-2.3 |
 | `LTX2ImageToVideoPipeline` | `LTX2Pipeline` with `image=` |
 | `LTX23ImageToVideoPipeline` | `LTX2Pipeline` with `image=`; checkpoint metadata selects LTX-2.3 |
-| `LTX2TwoStagesPipeline` | `LTX2DistilledTwoStagePipeline` |
-| `LTX2ImageToVideoTwoStagesPipeline` | `LTX2DistilledTwoStagePipeline` with `image=` |
+| `LTX2TwoStagesPipeline` | `LTX2DistilledPipeline` |
+| `LTX2ImageToVideoTwoStagesPipeline` | `LTX2DistilledPipeline` with `image=` |

 Passing any second positional argument now raises `TypeError`. These changes
 affect direct Python callers and explicit `--model-class-name` overrides;
@@ -117,10 +112,10 @@ Start a two-stage checkpoint with its explicit class:

 ```bash
 vllm serve rootonchair/LTX-2-19b-distilled --omni \
-  --model-class-name LTX2DistilledTwoStagePipeline --stage-init-timeout 600
+  --model-class-name LTX2DistilledPipeline --stage-init-timeout 600
 # LTX-2.3 full-distilled; the v1.1 x2 upsampler is resolved when absent
 vllm serve diffusers/LTX-2.3-Distilled-Diffusers --omni \
-  --model-class-name LTX2DistilledTwoStagePipeline --stage-init-timeout 600
+  --model-class-name LTX2DistilledPipeline --stage-init-timeout 600
 # Ordinary LTX-2.3 two-stage
 vllm serve diffusers/LTX-2.3-Diffusers --omni \
  --model-class-name LTX2TwoStagePipeline \
@@ -1,195 +0,0 @@
-# MiniMax-H3 on a single NPU 950PR
-
-This recipe runs MiniMax-H3 with online INT8 quantization on one NPU 950PR
-NPU (128 GB HiBL 1.0 HBM). It covers the single-card T2VA configuration at
-1024x576. For the eight-card Atlas 800I A3 BF16 route at 768P, see
-[MiniMax-H3-NPU.md](MiniMax-H3-NPU.md).
-
-## Capacity requirements
-
-| Resource | Requirement |
-| --- | ---: |
-| NPU | 1x NPU 950PR |
-| NPU HBM | 128 GiB (131,072 MiB reported by `npu-smi`) |
-| Observed HBM high-water mark | 118,442 MiB (90.4% of capacity) |
-| Checkpoint storage | 135 GiB per partition, local disk strongly preferred |
-| Container shared memory | 8 GiB, or a writable local filesystem for spill |
-
-The observed high-water mark leaves about 12.3 GiB of headroom, so this card is
-sized for exactly one resident task partition at 1024x576. Do not expect a
-larger output shape or concurrency greater than one to fit.
-
-Because a single 950PR holds the whole partition in HBM, this route does not
-need layerwise offload, and the 200 GiB system-RAM floor carried by the 72 GiB
-GPU recipes does not apply. The validated run completed inside a container
-limited to 32 GiB of RAM. Passing `--enable-layerwise-offload` on this card is
-actively harmful: it stages weights in non-reclaimable host memory and the
-container's OOM killer terminates the server with exit code -9.
-
-
-## Environment
-
- Host architecture: x86_64
-  driver: 25.7.rc1.6 (hal 7.35.23)
-  firmware: 9.0.0.105.229
- CANN toolkit: 9.1.0 (`/usr/local//cann-9.1.0`)
- npu-smi: 25.7.rc1.6
- Python: 3.12.13
- PyTorch: 2.10.0+cpu
- torch_npu: 2.10.0.post2
- vLLM: 0.26.0
- vLLM-Omni: 0.26.1.dev103+g584d78c67.npu (commit `584d78c6`)
-
-Install vLLM-Omni from a checkout with MiniMax-H3 support:
-
-```bash
-uv venv
-source .venv/bin/activate
-uv pip install -e .
-```
-
-## Start a server
-
-```bash
-export MODEL=/path/to/MiniMax-H3/FL2VA
-export PORT=8000
-export VLLM_WORKER_MULTIPROC_METHOD=spawn
-export VLLM_OMNI_VIDEO_SYNC_TIMEOUT=14400
-
-vllm serve "${MODEL}" \
-  --omni \
-  --host 0.0.0.0 \
-  --port "${PORT}" \
-  --trust-remote-code \
-  --num-gpus 1 \
-  --tensor-parallel-size 1 \
-  --usp 1 \
-  --ring 1 \
-  --text-encoder-tp-size 1 \
-  --vae-patch-parallel-size 1 \
-  --vae-parallel-mode tile \
-  --vae-use-tiling \
-  --init-timeout 14400 \
-  --stage-init-timeout 14400 \
-  --quantization int8
-```
-
-Do not pass `--diffusion-attention-backend CUDNN_ATTN`. cuDNN attention has no
-NPU implementation and fails at the first denoise step. Leave the backend
-unset: it resolves to `TORCH_SDPA`, or to `FLASH_ATTN` when MindIE-SD is
-installed (see [MiniMax-H3-NPU.md § Environment](MiniMax-H3-NPU.md#environment)).
-
-H3 is CFG-distilled, so `--cfg-parallel-size` must remain 1.
-
-## Validated evidence
-
-Measured on one NPU 950PR with the server command above plus
-`--enable-diffusion-pipeline-profiler`, generating a 5 s 1024x576 T2VA clip from
-the request in [§ T2VA request example](#t2va-request-example). MiniMax-H3
-requested 60 denoise steps and executed 59 denoise updates, so per-step latency
-is `denoise / 59`. `ffprobe` confirms every returned MP4 is 1024x576, 24 fps,
-124 video frames, 5.175 s, with a 32 kHz audio track of 165,600 samples.
-
-Five requests were issued back to back with no idle time between them. The card
-is measurably faster on the first request than in steady state, so both regimes
-are reported.
-
-| Stage | First request, cold card | Thermal steady state |
-| --- | ---: | ---: |
-| End-to-end request | 462.13 s | 506.38 s |
-| Text encode | 0.55 s | 0.05 s |
-| Denoise (59 updates) | 429.29 s | 471.74 s |
-| Per denoise update | 7,276 ms | 7,996 ms |
-| VAE decode | 8.43 s | 8.14 s |
-| Worker-to-server handoff | 24 s | 27 s |
-| MP4 muxing | 0.57 s | 0.50 s |
-| Server-reported `denoise_step_latency_ms` | 7,702 ms | 8,440 ms |
-| Peak HBM (`npu-smi`) | 118,442 MiB | 118,442 MiB |
-| Average NPU power | 566.9 W | 556.9 W |
-| Peak NPU temperature | 104 C | 104 C |
-
-Sample counts differ per row. The cold column is a single request. In the steady
-column, text encode, denoise and VAE decode are means over the four subsequent
-requests — denoise spanned 470.37 to 473.21 s, a 0.60% spread — while
-end-to-end, handoff and muxing are means over the two of those four that
-survived the handoff timeout described above.
-
-**Report the steady-state column.** The cold card completes denoise 9.9% faster
-while drawing only 1.8% more average power, and both regimes reach the same
-104 C ceiling, so the first request is buying a short window of higher clocks
-before the die saturates. Any benchmark that issues a single request against an
-idle card will overstate throughput by about 10%.
-
-**Do not read `denoise_step_latency_ms` as a per-step time.** The server divides
-the entire stage wall time by the 60 *requested* steps, so it absorbs text
-encoding, VAE decode and the 837 MiB handoff. All three successful requests
-satisfy `denoise_step_latency_ms == e2e_stage_wall_time_ms / 60` exactly. The
-per-update figure in the table is `diffuse / 59` from the pipeline profiler.
-
-Denoise accounts for 93% of the steady-state request and the handoff accounts
-for 5.3%, so keeping the transfer in shared memory should bring end-to-end down
-to roughly 480 s.
-
-Peak HBM is sampled externally with `npu-smi` at a 3 s interval. It is a
-reserved high-water mark rather than live allocation: it reads 111,272 MiB after
-weight loading and before the first request, and does not fall back after
-requests complete. Re-measure for longer outputs, a different output shape, or
-concurrency greater than one.
-
-### Startup
-
-Weight loading took **31 min 50 s** on the validated host (first log line
-08:04:59, `Application startup complete` 08:36:49), with the checkpoint on a
-shared GlusterFS network mount. This is why the server command sets
-`--init-timeout 14400` and `--stage-init-timeout 14400`; the stock 600 s and
-1800 s timeouts both abort mid-load with `TimeoutError` and exit code 143.
-Reduce both timeouts when the partition is staged on local disk.
-
-## T2VA request example
-
-```bash
-export API_URL="http://127.0.0.1:${PORT}/v1/videos/sync"
-
-curl -sS --max-time 1800 -X POST "${API_URL}" \
-  -F 'prompt=At night, three cats march into a bedroom playing tiny brass instruments, then abruptly file out, with synchronized room ambience.' \
-  -F 'width=1024' \
-  -F 'height=576' \
-  -F 'aspect_ratio=16:9' \
-  -F 'fps=24' \
-  -F 'num_inference_steps=60' \
-  -F 'flow_shift=12' \
-  -F 'seed=1101' \
-  -F 'extra_params={"task":"t2va","duration":5,"audio_flow_shift":3.0}' \
-  -o t2va.mp4
-```
-
-## Known limitations
-
- INT8 online quantization is validated for T2VA on this card. Use BF16 for
-  FL2VA and Ref2VA, or re-measure before relying on it.
- Single-card serving loads one task partition at a time. For Ref2VA, stop the
-  server and restart it against the `Ref2VA` directory.
- Sustained load drives the die to 104 C and costs about 10% of denoise
-  throughput relative to a cold card. Warm the server with at least one full
-  request before measuring, and treat steady state as the reportable number.
- Requests can fail during result handoff on hosts where the IPC spill path is
-  slow. See [§ Result handoff and shared
-  memory](#result-handoff-and-shared-memory).
- Loading the checkpoint from a network mount takes over 30 minutes and forces
-  very large init timeouts. Stage the partition on local disk when possible.
- The configuration measured here is 1024x576 at 60 steps, which differs from
-  the 1344x768 at 50 steps used by the GPU recipes in this directory. Denoise
-  cost scales superlinearly with token count, so the numbers are not directly
-  comparable across recipes.
- The image ships a `triton` package whose  backend is not built
-  (`No module named 'triton._C.libtriton.'`). vLLM logs this as an error
-  at startup and disables Triton; the diffusion path does not need it and the
-  run completes normally.
-
-## Additional resources
-
- [MiniMax-H3.md](MiniMax-H3.md) — full GPU guide
- [MiniMax-H3-NPU.md](MiniMax-H3-NPU.md) — eight-card Atlas 800I A3 BF16 guide
- [Int8 quantization](../../docs/user_guide/quantization/int8.md)
- [Supported models](../../docs/models/supported_models.md)
- [Video API](../../docs/serving/videos_api.md)
@@ -357,19 +357,12 @@ No restart is needed: `task=fl2va` routes to `FL2VA/transformer`, while

 ### Online FP8 quantization

-MiniMax H3 supports online FP8 quantization of both the DiT and the Qwen3-VL
-text decoder. The checkpoint remains BF16 on disk; vLLM-Omni creates FP8
-weights at runtime and uses dynamic activation scaling during inference. By
-default, `--quantization fp8` quantizes eligible attention and MLP linears in
-the text decoder, token refiner, and main DiT blocks, as well as the DiT
-condition and AdaLN projections. The Qwen vision tower, embeddings, norms,
-RoPE, both VAEs, and the model's FP32 patch, timestep, and output projections
-keep checkpoint precision.
-
-To select a component, use `--diffusion-quantization-config` with
-`{"transformer":{"method":"fp8"}}` for DiT-only FP8 or
-`{"text_encoder":{"method":"fp8"}}` for text-decoder-only FP8. The two
-entries can be combined. The shorthand below enables both components.
+MiniMax H3 supports load-time FP8 quantization of the DiT. The checkpoint
+remains BF16 on disk; vLLM-Omni quantizes eligible weights while loading and
+uses dynamic activation scaling during inference. By default, attention and
+MLP linears in the token refiner and main DiT blocks, the condition
+projection, and all AdaLN projections use FP8. Patch, timestep, and final
+projections remain FP32; the text encoder and VAEs are unchanged.

 Add this option to an existing H3 server command:

@@ -377,28 +370,6 @@ Add this option to an existing H3 server command:
 --quantization fp8
 ```

-#### Single 96 GB GPU, no-offload capacity check
-
-Use the FL2VA-only partition for this capacity test. Loading the combined
-service would also load the Ref2VA DiT and would test a different memory
-budget. A no-offload capacity check should contain none of
-`--enable-cpu-offload`, `--enable-layerwise-offload`, or
-`--enable-distributed-layerwise-offload`. VAE tiling changes decode placement
-but does not offload model weights to the CPU.
-
-The run passes the capacity check when the server initializes, the request
-finishes without CUDA OOM or Xid errors, `peak_used_mib` remains below the
-card's reported `memory_total_mib`, and `ffprobe` reports H.264 video plus
-32 kHz stereo AAC audio. Report the measured headroom rather than assuming
-that every nominal 96 GB SKU exposes the same MiB total.
-
-As a capacity proxy only, the same five-second case on one B300 measured a
-92,946 MiB whole-device first-request peak and a 92,146 MiB worker peak. Its
-encode, diffuse, decode, and client wall times were 8.664 s, 134.504 s,
-6.016 s, and 151.583 s. These numbers suggest that a 96 GB card may fit, but
-they are not an RTX PRO 6000 validation: kernels, allocator behavior, and
-reported device capacity differ across GPUs.
-
 Use `ignored_layers` to keep any otherwise eligible linear in BF16. H3
 resolves the `transformer` component before constructing the DiT, so names do
 not start with `transformer.`. Entries are exact runtime linear prefixes; a
@@ -415,10 +386,10 @@ For example, keep the first main block's attention projections in BF16 with:
  '{"transformer":{"method":"fp8","ignored_layers":["blocks.0.attn.qkv_proj","blocks.0.attn.out_proj"]}}'
 ```

-The structured option replaces `--quantization fp8`. Online FP8 can be used
-with H3 layerwise offload and with distributed layerwise offload's full-weight
-per-rank path (`--dlo-no-use-allgather`). The sharded DLO AllGather path is not
-supported for runtime-created FP8 weights.
+The structured option replaces `--quantization fp8`. Online FP8 is currently
+incompatible with H3 layerwise offload because the offload path produces a
+weight stride rejected by the Cutlass FP8 kernel. Use resident FP8 with tensor
+parallelism and VAE tiling instead.

 ## AMD ROCm (gfx942 / gfx950)

@@ -779,13 +750,6 @@ audio spectral cosine 0.9589 (minimum 0.80), and audio RMS ratio 0.9342. The
 resident per-GPU peak was 68.52 GiB for BF16 and 53.51 GiB for FP8, a 22%
 reduction.

-For direct human inspection, see the external
-[BF16 versus global-FP8 comparison](https://lishunyang12.github.io/vllm-omni-rankings/scripts/minimax_h3_global_fp8_vs_bf16/),
-which includes matched five-second T2VA, I2VA, and Ref2VA videos. The
-[comparison sources](https://github.com/lishunyang12/vllm-omni-rankings/tree/main/scripts/minimax_h3_global_fp8_vs_bf16)
-also record per-task fidelity metrics and provenance without storing generated
-media in this repository.
-
 ## TeaCache acceleration

 TeaCache reuses DiT block residuals across denoising steps when consecutive
@@ -870,8 +834,7 @@ vllm serve "${MODEL_ROOT}/FL2VA" \
  H3 native `tile` mode only.
 - A U2 x Ring2 hybrid currently fails with an attention-mask length mismatch; use
  pure Ulysses.
- Runtime-created FP8 weights do not support the sharded DLO AllGather path;
-  use `--dlo-no-use-allgather` when combining online FP8 with DLO.
+- Online FP8 is not compatible with layerwise offload.
 - TeaCache and Cache-DiT cannot be enabled on the same server.
 - Image+audio Ref2VA accepts exactly one image and one audio reference.
 - Video Ref2VA accepts one or more video files, but not an additional standalone
@@ -55,7 +55,6 @@ recipes/
 | [`MiniMaxAI/MiniMax-H3-MUSA.md`](./MiniMaxAI/MiniMax-H3-MUSA.md) | MiniMax H3 T2VA, FL2VA, and Ref2VA on MUSA | 4x MTT S5000 (TP4 Ref2VA profile) |
 | [`krea/Krea-2.md`](./krea/Krea-2.md) | Text-to-image (Turbo + Raw), offline + online, with LoRA | 1x H100 80GB |
 | [`LTX/LTX-2.md`](./LTX/LTX-2.md) | LTX-2/LTX-2.3 text-to-video and image-to-video with synchronized audio | H200 141GB / 96GB-class GPU |
-| [`LTX/LTX-2.5.md`](./LTX/LTX-2.5.md) | LTX-2.5-Diffusers: Full/SFT one-stage and distilled two-stage T2V/I2V with synchronized audio | NVIDIA B300; cuDNN-qualified |
 | [`MammothModa2/MammothModa2.md`](./MammothModa2/MammothModa2.md) | Preview and Dev text-to-image (AR → DiT); Dev text/image understanding | Preview: 1x L40S 48GB / 1x ≥40GB GPU; Dev: 1x NVIDIA GPU with sufficient cache headroom |
 | [`meituan-longcat/LongCat-Video-Avatar-1.5.md`](./meituan-longcat/LongCat-Video-Avatar-1.5.md) | Audio-driven avatar video generation (AT2V / AI2V, single- and multi-speaker, AVC continuation) | 1x H100 80GB |
 | [`mistralai/Voxtral-TTS.md`](./mistralai/Voxtral-TTS.md) | Online serving for TTS | 1x RTX 4090 24GB |
@@ -1,15 +1,15 @@
 # Common dependencies for all platforms
-# transformers: required unified multimodal model classes are unavailable
-# before 5.10.1; 5.10.0 is yanked, and 5.15 has a known construction
-# regression. transformers 5.9 also changed
-# _adjust_missing_and_unexpected_keys from
+# transformers: kept exactly aligned with the upstream vLLM constraint.
+# vLLM pins ">= 4.56.0, != 5.0.*, != 5.1.*, != 5.2.*, != 5.3.*, != 5.4.*, != 5.5.0".
+# The previous extra "< 5.9.0" exclusion was removed when re-aligning with
+# upstream: transformers 5.9 changed _adjust_missing_and_unexpected_keys from
 # list+list to set|set, breaking models that declare
 # _keys_to_ignore_on_load_unexpected as a list (e.g. OpenMOSS-Team/MOSS-TTS-Nano)
 # with "TypeError: unsupported operand type(s) for |: 'list' and 'set'".
 # That breakage is now handled in code instead of by pinning:
 # MossTTSNanoForGeneration wraps weight loading in
 # transformers_keys_to_ignore_compat (vllm_omni/model_executor/models/utils.py).
-transformers >= 5.10.1, < 5.15
+transformers >= 5.5.3
 av>=14.0.0
 omegaconf>=2.3.0
 diffusers==0.38.0
@@ -39,6 +39,8 @@ def _render(changed_files: list[str]) -> str:


 def test_bootstrap_if_injected_by_step_key() -> None:
+    import yaml
+
    rendered = _render_bootstrap_pipeline(
        BOOTSTRAP_STEPS_TEMPLATE,
        decision=resolve_ci_decision([]),
@@ -140,7 +142,12 @@ def test_mirror_hardwares_a2b3_npu_4_expands_agents_image_and_plugins() -> None:
    ]


-def _surviving_labels(doc: dict, changed_files: list[str]) -> set[str]:
+COVERAGE_PILOT_LABELS = ("Diffusion · Bagel Test", "TTS · Qwen3-TTS Base Test")
+MERGE_PIPELINE = Path(".buildkite/cuda/test-merge.yml")
+
+
+def _surviving_labels(changed_files: list[str]) -> set[str]:
+    doc = yaml.safe_load(MERGE_PIPELINE.read_text(encoding="utf-8"))
    rendered = _render_test_pipeline(doc, changed_files=changed_files)
    labels: set[str] = set()

@@ -156,29 +163,6 @@ def _surviving_labels(doc: dict, changed_files: list[str]) -> set[str]:
    return labels


-# Synthetic coverage-style job: shared inputs that change what the split measures.
-_COVERAGE_SHARED_INPUTS_DOC = {
-    "steps": [
-        {
-            "label": "Coverage Pilot",
-            "source_file_dependencies": [
-                "tests/e2e/online_serving/test_example.py",
-                ".buildkite/common/scripts/run_cov_split.sh",
-                "pyproject.toml",
-            ],
-            "commands": [".buildkite/common/scripts/run_cov_split.sh --model-id example"],
-        },
-        {
-            "label": "Unrelated Model Test",
-            "source_file_dependencies": [
-                "tests/e2e/online_serving/test_other.py",
-            ],
-            "commands": ["pytest -sv tests/e2e/online_serving/test_other.py"],
-        },
-    ],
-}
-
-
@pytest.mark.parametrize(
    "changed_file",
    [
@@ -186,17 +170,18 @@ _COVERAGE_SHARED_INPUTS_DOC = {
        "pyproject.toml",
    ],
 )
-def test_coverage_shared_inputs_select_dependent_job(changed_file: str) -> None:
-    """Jobs that list coverage shared inputs must stay selected when those files change."""
-    labels = _surviving_labels(_COVERAGE_SHARED_INPUTS_DOC, [changed_file])
-    assert "Coverage Pilot" in labels
-    assert "Unrelated Model Test" not in labels
+def test_coverage_pilots_selected_by_their_shared_inputs(changed_file: str) -> None:
+    """The split helper and the coverage config both change what these jobs measure.
+
+    Without them in source_file_dependencies the jobs are filtered out of normal
+    PR builds, so a regression would only surface in a later nightly.
+    """
+    labels = _surviving_labels([changed_file])
+    for pilot in COVERAGE_PILOT_LABELS:
+        assert pilot in labels, f"{pilot} dropped when {changed_file} changed"


-def test_coverage_shared_inputs_ignored_for_unrelated_change() -> None:
-    labels = _surviving_labels(
-        _COVERAGE_SHARED_INPUTS_DOC,
-        ["vllm_omni/entrypoints/openai/serving_chat.py"],
-    )
-    assert "Coverage Pilot" not in labels
-    assert "Unrelated Model Test" not in labels
+def test_coverage_pilots_not_selected_by_unrelated_change() -> None:
+    labels = _surviving_labels(["vllm_omni/entrypoints/openai/serving_chat.py"])
+    for pilot in COVERAGE_PILOT_LABELS:
+        assert pilot not in labels
@@ -14,8 +14,7 @@ import pytest
 from transformers import PretrainedConfig, Qwen3OmniMoeConfig

 from tests.helpers.stage_config import get_deploy_config_path, get_deploy_config_stage
-from vllm_omni.config import config_factory as config_factory_module
-from vllm_omni.config.config_factory import StageConfigFactory, _materialize_object_storage_configs
+from vllm_omni.config.config_factory import StageConfigFactory
 from vllm_omni.config.endpoint_policy import EndpointRestriction, OmniServingCapability
 from vllm_omni.config.omni_config import VllmOmniConfig
 from vllm_omni.config.pipeline_registry import OMNI_PIPELINES, register_pipeline, resolve_pipeline_config
@@ -56,7 +55,6 @@ def clear_config_factory_caches():
    yield
    StageConfigFactory.get_hf_config.cache_clear()
    StageConfigFactory.try_infer_model_type.cache_clear()
-    _materialize_object_storage_configs.cache_clear()


 Q3_OMNI_ALL_STAGES_HF_CONFIG = Qwen3OmniMoeConfig(enable_audio_output=True)
@@ -2767,89 +2765,3 @@ class TestPipelineConfigResolvers:
            pass

        assert resolver(NotTheRightHfConfig()) is None
-
-
-class TestObjectStorageConfigResolution:
-    """Object-storage URIs must be materialized locally before any HF-style reads.
-
-    Regression coverage for review on the Run:AI PR: ``get_config`` and the
-    config.json/model_index.json fallbacks crashed with ``HFValidationError``
-    for ``s3://`` URIs, and the name-match fallback scanned the whole URI, so
-    a bucket named after another pipeline could hijack pipeline selection.
-    """
-
-    @pytest.fixture
-    def fake_object_storage(self, monkeypatch, tmp_path):
-        """Replace ObjectStorageModel with a fake that "pulls" into a tmp dir.
-
-        Returns a recorder: calls to the returned callable write files into
-        the materialized directory, and ``recorder.pulls`` records every
-        ``pull_files`` invocation.
-        """
-        pulls: list[tuple[str, list[str] | None, list[str] | None]] = []
-        materialized_files: list[tuple[str, str]] = []
-
-        class FakeObjectStorageModel:
-            def __init__(self, url: str):
-                self.dir = str(tmp_path)
-
-            def pull_files(self, model_path, allow_pattern=None, ignore_pattern=None):
-                pulls.append((model_path, allow_pattern, ignore_pattern))
-                for name, content in materialized_files:
-                    (tmp_path / name).write_text(content)
-
-        monkeypatch.setattr(config_factory_module, "ObjectStorageModel", FakeObjectStorageModel)
-
-        class Recorder:
-            def set_files(self, files: list[tuple[str, str]]) -> None:
-                materialized_files.clear()
-                materialized_files.extend(files)
-
-        recorder = Recorder()
-        recorder.pulls = pulls
-        return recorder
-
-    def test_passthrough_for_non_uri(self, fake_object_storage):
-        assert _materialize_object_storage_configs("org/model") == "org/model"
-        assert _materialize_object_storage_configs("/local/model") == "/local/model"
-        assert fake_object_storage.pulls == []
-
-    def test_materialize_pulls_configs_once_per_uri(self, fake_object_storage):
-        uri = "s3://bucket/model"
-
-        first = _materialize_object_storage_configs(uri)
-        second = _materialize_object_storage_configs(uri)
-
-        assert first == second
-        assert len(fake_object_storage.pulls) == 1
-        pulled_uri, allow_pattern, ignore_pattern = fake_object_storage.pulls[0]
-        assert pulled_uri == uri
-        assert "*.json" in allow_pattern
-        assert ignore_pattern is None
-
-    def test_try_infer_model_type_reads_materialized_config(self, fake_object_storage):
-        """A misleading bucket name must not beat the real config.json content."""
-        fake_object_storage.set_files(
-            [
-                (
-                    "config.json",
-                    '{"model_type": "qwen3_omni_moe", "architectures": ["Qwen3OmniMoeForConditionalGeneration"]}',
-                )
-            ]
-        )
-        uri = "s3://qwen3-tts-models/Qwen3-Omni-30B-A3B-Instruct"
-
-        model = StageConfigFactory.try_infer_model_type(model=uri, trust_remote_code=False)
-
-        assert model == "qwen3_omni_moe"
-        assert fake_object_storage.pulls  # materialization happened
-
-    def test_name_match_fallback_scans_basename_only(self, fake_object_storage):
-        # Empty config.json (CosyVoice3 style) forces the name-match fallback.
-        fake_object_storage.set_files([("config.json", "{}")])
-
-        deceptive = "s3://qwen3-tts-models/plain-checkpoint"
-        assert StageConfigFactory.try_infer_model_type(model=deceptive, trust_remote_code=False) is None
-
-        matching = "s3://any-bucket/my-cosyvoice3-model"
-        assert StageConfigFactory.try_infer_model_type(model=matching, trust_remote_code=False) == "cosyvoice3"
@@ -424,10 +424,10 @@
        "baseline": {
          "H100": {
            "median_audio_ttfp_ms": [
-              82.6344
+              82.1886
            ],
            "median_audio_rtf": [
-              0.1277
+              0.1241
            ]
          }
        }
@@ -453,13 +453,13 @@
        "baseline": {
          "H100": {
            "median_audio_ttfp_ms": [
-              170.3737
+              160.9168
            ],
            "median_audio_rtf": [
-              0.2076
+              0.1907
            ],
            "audio_throughput": [
-              37.0834
+              39.2332
            ]
          }
        }
@@ -485,13 +485,13 @@
        "baseline": {
          "H100": {
            "median_audio_ttfp_ms": [
-              101.0293
+              100.1265
            ],
            "median_audio_rtf": [
-              0.1498
+              0.1437
            ],
            "audio_throughput": [
-              10.1186
+              11.7605
            ]
          }
        }
@@ -5,7 +5,6 @@ from contextlib import nullcontext

 import pytest
 import torch
-from torch.nn.attention import SDPBackend

 from vllm_omni.diffusion.attention.backends.abstract import AttentionMetadata
 from vllm_omni.diffusion.attention.backends.cudnn_attn import CuDNNAttentionImpl
@@ -13,48 +12,6 @@ from vllm_omni.diffusion.attention.backends.cudnn_attn import CuDNNAttentionImpl
 pytestmark = [pytest.mark.diffusion, pytest.mark.cpu, pytest.mark.core_model]


-@pytest.mark.parametrize(
-    ("query_sequence_length", "key_sequence_length", "expected_backend"),
-    [
-        (1, 1, SDPBackend.MATH),
-        (8, 1, SDPBackend.MATH),
-        (1, 8, SDPBackend.MATH),
-        (8, 8, SDPBackend.CUDNN_ATTENTION),
-    ],
-)
-def test_cudnn_routes_single_token_attention_to_math(
-    monkeypatch, query_sequence_length, key_sequence_length, expected_backend
-):
-    selected_backends = []
-
-    def fake_sdpa_kernel(backends):
-        selected_backends.append(backends)
-        return nullcontext()
-
-    monkeypatch.setattr(
-        "vllm_omni.diffusion.attention.backends.cudnn_attn.sdpa_kernel",
-        fake_sdpa_kernel,
-    )
-    monkeypatch.setattr(
-        torch.nn.functional,
-        "scaled_dot_product_attention",
-        lambda query, _key, _value, **_kwargs: query,
-    )
-    impl = CuDNNAttentionImpl(
-        num_heads=2,
-        head_size=4,
-        softmax_scale=0.5,
-    )
-    query = torch.randn(1, query_sequence_length, 2, 4)
-    key = torch.randn(1, key_sequence_length, 2, 4)
-    value = torch.randn_like(key)
-
-    output = impl.forward_cuda(query, key, value)
-
-    assert output.shape == query.shape
-    assert selected_backends == [[expected_backend]]
-
-
 def test_cudnn_slices_valid_kv_prefix_without_padding_mask(monkeypatch):
    observed = {}

@@ -36,6 +36,7 @@ pytestmark = [pytest.mark.core_model, pytest.mark.cpu]

 SEPARATE_CFG_TRANSFORMERS = [
    DreamIdOmniModel,
+    LTX2VideoTransformer3DModel,
    HeliosTransformer3DModel,
    LongCatImageTransformer2DModel,
    Cosmos3VFMTransformer,
@@ -123,13 +124,6 @@ def test_cache_dit_configs_have_separate_cfg(transformer_model):
    assert transformer_model._cache_dit_adapter_config.has_separate_cfg is True


-def test_ltx2_cache_dit_uses_one_forward_per_denoise_step():
-    """LTX batches all guidance passes into one Transformer invocation."""
-    adapter_config = LTX2VideoTransformer3DModel._cache_dit_adapter_config
-
-    assert adapter_config.has_separate_cfg is False
-
-
@patch("vllm_omni.diffusion.cache.cachedit.model_specific.BlockAdapter")
@patch("vllm_omni.diffusion.cache.cachedit.model_specific.cache_dit")
 def test_separate_wan22_custom_enabler_has_separate_cfg(mock_cache_dit, mock_block_adapter):
@@ -53,10 +53,14 @@ def _diff_metrics(a: Image.Image, b: Image.Image) -> tuple[float, float]:
 def _extract_single_image(outputs) -> Image.Image:
    first_output = outputs[0]
    assert first_output.final_output_type == "image"
-    if not isinstance(first_output, OmniRequestOutput):
-        raise ValueError(f"Expected an OmniRequestOutput, got {type(first_output).__name__}")
+    if not hasattr(first_output, "request_output") or not first_output.request_output:
+        raise ValueError("No request_output found in OmniRequestOutput")

-    images = first_output.images
+    req_out = first_output.request_output
+    if not isinstance(req_out, OmniRequestOutput) or not hasattr(req_out, "images"):
+        raise ValueError("Invalid request_output structure or missing 'images' key")
+
+    images = req_out.images
    if images is None or len(images) != 1:
        raise ValueError(f"Expected 1 image, got {0 if images is None else len(images)}")
    return images[0]
@@ -72,7 +72,8 @@ def _to_float_array(data: Any) -> np.ndarray:


 def _extract_output_array(output: Any) -> np.ndarray:
-    payload = output.images[0]
+    req_out = output.request_output
+    payload = req_out.images[0]
    return _to_float_array(payload)


@@ -16,11 +16,6 @@ from vllm.config.load import LoadConfig
 from vllm_omni.diffusion.config import get_current_diffusion_config, get_current_diffusion_config_or_none
 from vllm_omni.diffusion.data import OmniDiffusionConfig
 from vllm_omni.diffusion.model_loader.diffusers_loader import DiffusersPipelineLoader
-from vllm_omni.diffusion.model_loader.host_weight_plan import (
-    HostWeightPlan,
-    HostWeightPlanResult,
-    TensorBinding,
-)
 from vllm_omni.diffusion.models.helios import HeliosPipeline
 from vllm_omni.diffusion.registry import initialize_model

@@ -115,85 +110,6 @@ def test_empty_source_prefix_keeps_full_model_strict_check():
        loader.load_weights(model)


-def test_stream_online_quant_weights_offloads_layers_after_processing():
-    from vllm.model_executor.model_loader.reload.layerwise import (
-        get_layerwise_info,
-    )
-
-    events: list[str] = []
-
-    class _OnlineQuantMethod:
-        uses_meta_device = True
-
-    class _TrackedLayer(nn.Linear):
-        def __init__(self, name: str):
-            super().__init__(2, 2, bias=False)
-            self.name = name
-            self.quant_method = _OnlineQuantMethod()
-            get_layerwise_info(self).load_numel_total = self.weight.numel()
-
-        def to(self, *args, **kwargs):
-            events.append(self.name)
-            return super().to(*args, **kwargs)
-
-    class _StreamingModel(nn.Module):
-        def __init__(self):
-            super().__init__()
-            self.first = _TrackedLayer("first")
-            self.second = _TrackedLayer("second")
-
-    model = _StreamingModel()
-    weights = iter(
-        [
-            ("first.weight", torch.zeros((2, 2))),
-            ("second.weight", torch.zeros((2, 2))),
-        ]
-    )
-    streamed = DiffusersPipelineLoader._stream_online_quant_weights_to_cpu(model, weights)
-
-    assert next(streamed)[0] == "first.weight"
-    get_layerwise_info(model.first).reset()
-    assert next(streamed)[0] == "second.weight"
-    assert events == ["first"]
-
-    get_layerwise_info(model.second).reset()
-    with pytest.raises(StopIteration):
-        next(streamed)
-    assert events == ["first", "second"]
-
-
-def test_process_weights_skips_completed_online_quant_layer(monkeypatch):
-    from unittest.mock import Mock
-
-    from vllm.model_executor.layers.quantization.base_config import QuantizeMethodBase
-    from vllm.model_executor.model_loader.reload import layerwise
-
-    class _TrackedLayer(nn.Linear):
-        def __init__(self):
-            super().__init__(2, 2, bias=False)
-            self.to_calls: list[object] = []
-
-        def to(self, *args, **kwargs):
-            self.to_calls.append(args[0] if args else kwargs.get("device"))
-            return super().to(*args, **kwargs)
-
-    model = nn.Module()
-    model.layer = _TrackedLayer()
-    quant_method = Mock(spec=QuantizeMethodBase)
-    quant_method.uses_meta_device = True
-    model.layer.quant_method = quant_method
-    model.layer._already_called_process_weights_after_loading = True
-    finalize = Mock()
-    monkeypatch.setattr(layerwise, "finalize_layerwise_processing", finalize)
-
-    loader = _make_loader_with_weights([])
-    loader._process_weights_after_loading(model, torch.device("cuda"))
-
-    finalize.assert_called_once_with(model, model_config=None)
-    quant_method.process_weights_after_loading.assert_not_called()
-    assert model.layer.to_calls == []
-
-
 class _ConfigAwareModel(nn.Module):
    def __init__(self, *, od_config):
        super().__init__()
@@ -263,137 +179,6 @@ def test_load_model_custom_pipeline_sets_current_diffusion_config(monkeypatch):
    assert get_current_diffusion_config_or_none() is None


-def test_dlo_transfers_loader_plan_and_skips_ordinary_weight_loading(monkeypatch):
-    import vllm_omni.diffusion.model_loader.diffusers_loader as loader_mod
-
-    od_config = SimpleNamespace(
-        dtype=torch.float32,
-        parallel_config=SimpleNamespace(use_hsdp=False, tensor_parallel_size=1),
-        quantization_config=None,
-        enable_distributed_layerwise_offload=True,
-        dlo_use_allgather=False,
-        model="unused",
-    )
-    loader = DiffusersPipelineLoader(LoadConfig(), od_config)
-    model = nn.Module()
-    model.transformer = nn.Linear(2, 2, bias=False)
-    plan = HostWeightPlan(
-        backing_kind="checkpoint_mmap",
-        bindings={},
-    )
-    loaded_ordinary_weights = False
-
-    def load_weights(_model):
-        nonlocal loaded_ordinary_weights
-        loaded_ordinary_weights = True
-
-    loader._init_from_load_format = lambda *_args, **_kwargs: model  # type: ignore[method-assign]
-    loader.load_weights = load_weights  # type: ignore[method-assign]
-    loader._apply_skip_softmax_calibration = lambda _model: None  # type: ignore[method-assign]
-    monkeypatch.setattr(
-        loader_mod,
-        "build_checkpoint_mmap_plan",
-        lambda *_args, **_kwargs: HostWeightPlanResult(plan),
-    )
-
-    assert loader.load_model(load_device="cpu") is model
-    assert not loaded_ordinary_weights
-    assert loader.take_host_weight_plan() is plan
-    assert loader.take_host_weight_plan() is None
-
-
-def test_dlo_plan_loads_component_sources_outside_planned_dit(monkeypatch):
-    import vllm_omni.diffusion.model_loader.diffusers_loader as loader_mod
-
-    od_config = SimpleNamespace(
-        dtype=torch.float32,
-        parallel_config=SimpleNamespace(use_hsdp=False, tensor_parallel_size=1),
-        quantization_config=None,
-        enable_distributed_layerwise_offload=True,
-        dlo_use_allgather=False,
-        model="unused",
-    )
-    loader = DiffusersPipelineLoader(LoadConfig(), od_config)
-
-    class MixedSourceModel(nn.Module):
-        def __init__(self):
-            super().__init__()
-            self.transformer = nn.Linear(2, 2, bias=False)
-            self.text_encoder = nn.Linear(2, 2, bias=False)
-            self.weights_sources = (
-                DiffusersPipelineLoader.ComponentSource("unused", None, None, prefix="transformer."),
-                DiffusersPipelineLoader.ComponentSource("unused", None, None, prefix="text_encoder."),
-            )
-            self.loaded_weight_names: list[str] = []
-
-        def load_weights(self, weights):
-            self.loaded_weight_names = [name for name, _ in weights]
-            return set(self.loaded_weight_names)
-
-    model = MixedSourceModel()
-    plan = HostWeightPlan(
-        backing_kind="checkpoint_mmap",
-        bindings={
-            "transformer.weight": TensorBinding(
-                checkpoint_key="weight",
-                file_path="unused",
-            )
-        },
-        planned_source_prefixes=frozenset({"transformer."}),
-    )
-    requested_prefixes: list[str] = []
-
-    def get_weights(source, model=None):
-        del model
-        requested_prefixes.append(source.prefix)
-        yield source.prefix + "weight", torch.ones(2, 2)
-
-    loader._init_from_load_format = lambda *_args, **_kwargs: model  # type: ignore[method-assign]
-    loader._get_weights_iterator = get_weights  # type: ignore[method-assign]
-    loader._apply_skip_softmax_calibration = lambda _model: None  # type: ignore[method-assign]
-    monkeypatch.setattr(
-        loader_mod,
-        "build_checkpoint_mmap_plan",
-        lambda *_args, **_kwargs: HostWeightPlanResult(plan),
-    )
-
-    assert loader.load_model(load_device="cpu") is model
-    assert requested_prefixes == ["text_encoder."]
-    assert model.loaded_weight_names == ["text_encoder.weight"]
-    assert loader.take_host_weight_plan() is plan
-
-
-def test_dlo_plan_fallback_runs_ordinary_loader(monkeypatch):
-    import vllm_omni.diffusion.model_loader.diffusers_loader as loader_mod
-
-    od_config = SimpleNamespace(
-        dtype=torch.float32,
-        parallel_config=SimpleNamespace(use_hsdp=False, tensor_parallel_size=1),
-        quantization_config=None,
-        enable_distributed_layerwise_offload=True,
-        dlo_use_allgather=False,
-        model="unused",
-    )
-    loader = DiffusersPipelineLoader(LoadConfig(), od_config)
-    model = nn.Module()
-    model.transformer = nn.Linear(2, 2, bias=False)
-    calls: list[str] = []
-
-    loader._init_from_load_format = lambda *_args, **_kwargs: model  # type: ignore[method-assign]
-    loader.load_weights = lambda _model: calls.append("load")  # type: ignore[method-assign]
-    loader._process_weights_after_loading = lambda *_args: calls.append("process")  # type: ignore[method-assign]
-    loader._apply_skip_softmax_calibration = lambda _model: None  # type: ignore[method-assign]
-    monkeypatch.setattr(
-        loader_mod,
-        "build_checkpoint_mmap_plan",
-        lambda *_args, **_kwargs: HostWeightPlanResult(None, "not direct-compatible"),
-    )
-
-    assert loader.load_model(load_device="cpu") is model
-    assert calls == ["load", "process"]
-    assert loader.take_host_weight_plan() is None
-
-
 def test_hsdp_processes_quantized_weights_before_sharding(mocker):
    import vllm_omni.diffusion.model_loader.diffusers_loader as loader_mod
    from vllm_omni.diffusion.offloader.module_collector import PipelineModules
@@ -1,72 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-
-import contextlib
-
-import huggingface_hub
-import pytest
-
-from vllm_omni.diffusion.model_loader import hub_prefetch
-
-pytestmark = [pytest.mark.core_model, pytest.mark.diffusion, pytest.mark.cpu]
-
-
-def test_prefetch_subfolders_propagates_revision(monkeypatch):
-    calls = []
-
-    def fake_snapshot_download(**kwargs):
-        calls.append(kwargs)
-
-    monkeypatch.setattr(huggingface_hub, "snapshot_download", fake_snapshot_download)
-    monkeypatch.setattr(hub_prefetch, "_repo_prefetch_lock", lambda _model: contextlib.nullcontext())
-
-    hub_prefetch.prefetch_subfolders(
-        "org/model",
-        ["transformer"],
-        revision="pinned-revision",
-    )
-
-    assert calls == [
-        {
-            "repo_id": "org/model",
-            "revision": "pinned-revision",
-            "allow_patterns": ["transformer/*", "transformer/**", "*.json", "*.txt"],
-        }
-    ]
-
-
-def test_from_pretrained_retry_preserves_revision(monkeypatch):
-    factory_calls = []
-    prefetch_calls = []
-
-    def fake_factory(model, **kwargs):
-        factory_calls.append((model, kwargs))
-        if len(factory_calls) == 1:
-            raise OSError("partial cache")
-        return "loaded"
-
-    def fake_prefetch(model, subfolders, **kwargs):
-        prefetch_calls.append((model, tuple(subfolders), kwargs))
-
-    monkeypatch.setattr(hub_prefetch, "prefetch_subfolders", fake_prefetch)
-    monkeypatch.setattr(hub_prefetch.time, "sleep", lambda _seconds: None)
-
-    result = hub_prefetch.from_pretrained_with_prefetch(
-        fake_factory,
-        "org/model",
-        subfolder="transformer",
-        prefetch_list=["transformer", "vae"],
-        revision="pinned-revision",
-        max_attempts=2,
-    )
-
-    assert result == "loaded"
-    assert len(factory_calls) == 2
-    assert all(call[1]["revision"] == "pinned-revision" for call in factory_calls)
-    assert prefetch_calls == [
-        (
-            "org/model",
-            ("transformer", "vae"),
-            {"local_files_only": False, "revision": "pinned-revision"},
-        )
-    ]
@@ -7,17 +7,17 @@ from vllm_omni.diffusion.models.flux2.pipeline_flux2 import (
    Flux2Pipeline,
    _resolve_component_quant_config,
 )
-from vllm_omni.quantization import ComponentQuantizationConfig

 pytestmark = [pytest.mark.core_model, pytest.mark.cpu]


 def test_resolve_component_quant_config_routes_components():
    text_encoder_config = object()
-    component_config = ComponentQuantizationConfig(
-        component_configs={"text_encoder": text_encoder_config},
-        default_config=None,
-    )
+    component_config = Mock()
+    component_config.resolve.side_effect = {
+        "text_encoder": text_encoder_config,
+        "transformer": None,
+    }.get

    assert _resolve_component_quant_config(component_config, "text_encoder") is text_encoder_config
    assert _resolve_component_quant_config(component_config, "transformer") is None
@@ -1,881 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-
-"""Focused LTX-2.5 pipeline correctness tests."""
-
-import json
-from dataclasses import replace
-from types import SimpleNamespace
-
-import pytest
-import torch
-
-from vllm_omni.diffusion.data import DiffusionOutput
-from vllm_omni.diffusion.models.ltx2 import ltx2_components, ltx2_latents, ltx2_runtime
-from vllm_omni.diffusion.models.ltx2.ltx2_components import (
-    LTX25_DISTILLED_COMPONENT_PROFILE,
-    LTX25_DISTILLED_ONE_STAGE_COMPONENT_PROFILE,
-    LTX25_FULL_COMPONENT_PROFILE,
-    LTX25_TWO_STAGE_COMPONENT_PROFILE,
-    detect_ltx_model_version,
-    resolve_ltx_checkpoint_kind,
-    resolve_ltx_component_profile,
-)
-from vllm_omni.diffusion.models.ltx2.ltx2_denoise import (
-    LTXPhaseResult,
-    _first_frame_keyframes_mask,
-    _official_ltx_sigmas,
-)
-from vllm_omni.diffusion.models.ltx2.ltx2_guidance import (
-    LTX_GUIDANCE_EXECUTOR,
-    LTXGuidanceSpec,
-    LTXModalityGuidance,
-    velocity_from_x0,
-    x0_from_velocity,
-)
-from vllm_omni.diffusion.models.ltx2.ltx2_latents import LTXAVState
-from vllm_omni.diffusion.models.ltx2.ltx2_recipes import (
-    LTX2_DISTILLED_TWO_STAGE_RECIPE,
-    LTX2_ONE_STAGE_RECIPE,
-    LTX23_ONE_STAGE_RECIPE,
-    LTX25_DISTILLED_ONE_STAGE_RECIPE,
-    LTX25_DISTILLED_TWO_STAGE_RECIPE,
-    LTX25_FULL_RECIPE,
-    LTX25_TWO_STAGE_RECIPE,
-    LTX_DISTILLED_SIGMAS,
-    LTX_POSITIVE_ONLY_RECIPE,
-    resolve_ltx_pipeline_recipe,
-)
-from vllm_omni.diffusion.models.ltx2.ltx2_request import LTXRequestInputs
-from vllm_omni.diffusion.models.ltx2.ltx2_runtime import LTXRuntime
-from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import (
-    LTX2DistilledOneStagePipeline,
-    LTX2Pipeline,
-)
-from vllm_omni.diffusion.models.ltx2.pipeline_ltx2_two_stage import (
-    LTX2DistilledTwoStagePipeline,
-    LTX2TwoStagePipeline,
-)
-
-pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
-
-
-@pytest.mark.parametrize("model_version", ["2", "2.3", "2.5"])
-def test_ltx_timestep_adaln_expands_per_token_like_official_without_sp(model_version):
-    pipe = object.__new__(LTX2Pipeline)
-    pipe.model_version = model_version
-    pipe.od_config = SimpleNamespace(parallel_config=SimpleNamespace(sequence_parallel_size=1))
-    ts = torch.tensor([750.0, 500.0])
-
-    kwargs = LTXRuntime._denoise_timestep_kwargs(
-        pipe,
-        ts,
-        SimpleNamespace(),
-        SimpleNamespace(),
-        video_token_count=4,
-        audio_token_count=3,
-    )
-
-    torch.testing.assert_close(kwargs["timestep"], ts[:, None].expand(-1, 4))
-    torch.testing.assert_close(kwargs["audio_timestep"], ts[:, None].expand(-1, 3))
-    assert kwargs["sigma"] is ts
-    assert kwargs["audio_sigma"] is ts
-
-
-def test_ltx_ancestral_step_clears_sp_audio_padding_after_scheduler_update(monkeypatch):
-    state = LTXAVState(
-        video=torch.zeros(1, 2, 2),
-        audio=torch.zeros(1, 4, 2),
-    )
-    forward_ctx = SimpleNamespace(
-        original_audio_num_frames=3,
-        sampler="euler_ancestral",
-    )
-    denoise_ctx = SimpleNamespace(latents=None, audio_latents=None)
-    pipe = SimpleNamespace(
-        _predict_noise_for_step=lambda *_args: (torch.zeros_like(state.video), torch.zeros_like(state.audio))
-    )
-
-    def fake_step(_pipeline, actual_forward_ctx, _denoise_ctx, *_args):
-        assert actual_forward_ctx.sampler == "euler_ancestral"
-        updated_audio = torch.ones_like(state.audio)
-        updated_audio[:, 3:] = 99.0
-        return torch.ones_like(state.video), updated_audio
-
-    monkeypatch.setattr(ltx2_runtime, "step_denoised_latents", fake_step)
-
-    actual = LTXRuntime._denoise_step(
-        pipe,
-        0,
-        torch.tensor(1.0),
-        state,
-        forward_ctx,
-        denoise_ctx,
-    )
-
-    torch.testing.assert_close(actual.audio[:, :3], torch.ones_like(actual.audio[:, :3]))
-    torch.testing.assert_close(actual.audio[:, 3:], torch.zeros_like(actual.audio[:, 3:]))
-
-
-def test_ltx25_first_causal_latent_frame_is_marked_for_keyframe_embedding():
-    reference = torch.zeros(2, 12, 8)
-
-    mask = _first_frame_keyframes_mask(reference, latent_num_frames=3)
-
-    assert mask.shape == (2, 12, 1)
-    torch.testing.assert_close(mask[:, :4], torch.ones(2, 4, 1))
-    torch.testing.assert_close(mask[:, 4:], torch.zeros(2, 8, 1))
-
-
-def _make_ltx_request_pipe(cls):
-    pipe = object.__new__(cls)
-    torch.nn.Module.__init__(pipe)
-    pipe.device = torch.device("cpu")
-    pipe.tokenizer_max_length = 99
-    pipe.vae_spatial_compression_ratio = 32
-    pipe.vae_temporal_compression_ratio = 8
-    return pipe
-
-
-def _resolve_request_inputs_for_test(
-    pipe,
-    req,
-    *,
-    guidance_scale=4.0,
-    negative_prompt=None,
-):
-    return pipe._resolve_request_inputs(
-        req,
-        prompt=None,
-        negative_prompt=negative_prompt,
-        height=None,
-        width=None,
-        num_frames=None,
-        frame_rate=None,
-        num_inference_steps=None,
-        guidance_scale=guidance_scale,
-        num_videos_per_prompt=1,
-        generator=None,
-        latents=None,
-        audio_latents=None,
-        prompt_embeds=None,
-        negative_prompt_embeds=None,
-        prompt_attention_mask=None,
-        negative_prompt_attention_mask=None,
-        decode_timestep=0.0,
-        decode_noise_scale=None,
-        output_type="np",
-        max_sequence_length=None,
-    )
-
-
-def test_ltx25_missing_gemma4_recommends_supported_transformers_range(monkeypatch):
-    pipe = SimpleNamespace(component_profile=replace(LTX25_FULL_COMPONENT_PROFILE, text_encoder_cls=None))
-    od_config = SimpleNamespace(model="Lightricks/LTX-2.5-Diffusers", dtype=torch.bfloat16)
-
-    monkeypatch.setattr(ltx2_components, "get_local_device", lambda: torch.device("cpu"))
-    monkeypatch.setattr(ltx2_components, "prefetch_subfolders", lambda *_args, **_kwargs: None)
-    monkeypatch.setattr(ltx2_components.AutoTokenizer, "from_pretrained", lambda *_args, **_kwargs: object())
-
-    with pytest.raises(ImportError) as exc_info:
-        ltx2_components.initialize_pipeline_components(pipe, od_config)
-
-    assert str(exc_info.value) == (
-        "LTX-2.5 requires Gemma4UnifiedForConditionalGeneration; install transformers>=5.10.1,<5.15."
-    )
-
-
-def test_ltx_converted_component_loading_propagates_revision(monkeypatch):
-    revision = "pinned-revision"
-    calls = {"components": []}
-    profile = replace(
-        LTX25_DISTILLED_COMPONENT_PROFILE,
-        text_encoder_cls=object,
-        video_vae_cls=object,
-        vocoder_cls=object,
-        vocoder_fallback_cls=None,
-        scheduler_use_dynamic_shifting=False,
-    )
-    pipeline = SimpleNamespace(component_profile=profile, pipeline_kind="distilled_two_stage")
-    od_config = SimpleNamespace(
-        model="org/converted-ltx25",
-        revision=revision,
-        dtype=torch.bfloat16,
-        quantization_config=None,
-    )
-
-    monkeypatch.setattr(ltx2_components, "get_local_device", lambda: torch.device("cpu"))
-    monkeypatch.setattr(ltx2_components, "_install_connector_attention", lambda _connectors, **_kwargs: None)
-    monkeypatch.setattr(ltx2_components, "_place_aux_components", lambda _pipeline: None)
-
-    def fake_prefetch(model, subfolders, **kwargs):
-        calls["prefetch"] = (model, tuple(subfolders), kwargs)
-
-    def fake_tokenizer(*args, **kwargs):
-        calls["tokenizer"] = (args, kwargs)
-        return SimpleNamespace(model_max_length=1024)
-
-    def fake_component(component_cls, model, subfolder, **kwargs):
-        calls["components"].append((component_cls, model, subfolder, kwargs))
-        if subfolder == "vae":
-            return SimpleNamespace(spatial_compression_ratio=32, temporal_compression_ratio=8)
-        if subfolder == "audio_vae":
-            return SimpleNamespace(
-                mel_compression_ratio=4,
-                temporal_compression_ratio=4,
-                config=SimpleNamespace(sample_rate=16_000, mel_hop_length=160),
-            )
-        return object()
-
-    def fake_transformer_config(model, subfolder, local_files_only, *, revision):
-        calls["transformer_config"] = (model, subfolder, local_files_only, revision)
-        return {"component": "transformer"}
-
-    monkeypatch.setattr(ltx2_components, "prefetch_subfolders", fake_prefetch)
-    monkeypatch.setattr(ltx2_components.AutoTokenizer, "from_pretrained", fake_tokenizer)
-    monkeypatch.setattr(ltx2_components, "_load_component", fake_component)
-    monkeypatch.setattr(ltx2_components, "load_transformer_config", fake_transformer_config)
-    monkeypatch.setattr(
-        ltx2_components,
-        "create_transformer_from_config",
-        lambda *_args, **_kwargs: SimpleNamespace(config=SimpleNamespace(patch_size=1, patch_size_t=1)),
-    )
-
-    def fake_scheduler(*args, **kwargs):
-        calls["scheduler"] = (args, kwargs)
-        return SimpleNamespace(config={"use_dynamic_shifting": False, "shift_terminal": None})
-
-    monkeypatch.setattr(ltx2_components.FlowMatchEulerDiscreteScheduler, "from_pretrained", fake_scheduler)
-
-    ltx2_components.initialize_pipeline_components(pipeline, od_config)
-
-    assert pipeline.weights_sources[0].revision == revision
-    assert calls["prefetch"][2]["revision"] == revision
-    assert calls["tokenizer"][1]["revision"] == revision
-    assert {call[2] for call in calls["components"]} == {
-        "text_encoder",
-        "connectors",
-        "vae",
-        "audio_vae",
-        "vocoder",
-        "latent_upsampler",
-    }
-    assert all(call[3]["revision"] == revision for call in calls["components"])
-    assert calls["transformer_config"] == (
-        od_config.model,
-        profile.transformer_subfolder,
-        False,
-        revision,
-    )
-    assert calls["scheduler"][1]["revision"] == revision
-
-
-def test_ltx_checkpoint_explicit_version_precedes_structural_heuristics(tmp_path):
-    (tmp_path / "model_index.json").write_text(json.dumps({"model_version": "2.3.1"}))
-    (tmp_path / "transformer").mkdir()
-    (tmp_path / "transformer" / "config.json").write_text(json.dumps({"ff_bias": False}))
-
-    assert detect_ltx_model_version(str(tmp_path)) == "2.3"
-
-
-def test_ltx25_checkpoint_selects_full_one_stage_profile(tmp_path, monkeypatch):
-    from vllm_omni.diffusion.models.ltx2 import ltx2_runtime
-
-    (tmp_path / "model_index.json").write_text(
-        json.dumps({"text_encoder": ["transformers", "Gemma4UnifiedForConditionalGeneration"]})
-    )
-
-    def stub_components(pipe, od_config):
-        pipe.od_config = od_config
-        pipe.vae_spatial_compression_ratio = 32
-
-    monkeypatch.setattr(ltx2_runtime, "initialize_pipeline_components", stub_components)
-    monkeypatch.setattr(LTXRuntime, "setup_diffusion_pipeline_profiler", lambda *_args, **_kwargs: None)
-
-    pipe = LTX2Pipeline(od_config=SimpleNamespace(model=str(tmp_path), enable_diffusion_pipeline_profiler=False))
-
-    assert pipe.model_version == "2.5"
-    assert pipe.component_profile is LTX25_FULL_COMPONENT_PROFILE
-    assert pipe.pipeline_recipe is LTX25_FULL_RECIPE
-    assert pipe.preserve_sp_padded_audio_duration
-    assert pipe.reports_stage_durations
-
-
-def test_ltx25_full_recipe_matches_official_sft_defaults():
-    (phase,) = LTX25_FULL_RECIPE.phases
-
-    assert (LTX25_FULL_RECIPE.height, LTX25_FULL_RECIPE.width) == (544, 960)
-    assert LTX25_FULL_RECIPE.num_frames == 121
-    assert LTX25_FULL_RECIPE.frame_rate == 24.0
-    assert LTX25_FULL_RECIPE.num_inference_steps == 30
-    assert phase.guidance.video == LTXModalityGuidance(
-        cfg_scale=3.0,
-        stg_scale=1.0,
-        modality_scale=3.0,
-        rescale_scale=0.7,
-        stg_blocks=(28,),
-    )
-    assert phase.guidance.audio == LTXModalityGuidance(
-        cfg_scale=7.0,
-        stg_scale=1.0,
-        modality_scale=3.0,
-        rescale_scale=0.7,
-        stg_blocks=(28,),
-    )
-    assert phase.sigmas is None
-    assert phase.noise_scale == 1.0
-    assert phase.use_official_sigma_schedule
-    assert LTX25_FULL_COMPONENT_PROFILE.transformer_subfolder == "transformer_full"
-    assert LTX25_FULL_COMPONENT_PROFILE.scheduler_use_dynamic_shifting
-    assert LTX25_FULL_COMPONENT_PROFILE.scheduler_shift_terminal == 0.1
-
-
-def test_ltx25_full_sigma_schedule_uses_official_default_sequence_anchor():
-    scheduler = SimpleNamespace(
-        config={
-            "base_image_seq_len": 1024,
-            "max_image_seq_len": 4096,
-            "base_shift": 0.95,
-            "max_shift": 2.05,
-            "shift_terminal": 0.1,
-        }
-    )
-
-    sigmas = _official_ltx_sigmas(
-        scheduler,
-        steps=30,
-        device=torch.device("cpu"),
-    )
-
-    torch.testing.assert_close(
-        sigmas[[0, 1, -2, -1]],
-        torch.tensor([1.0, 0.9949570, 0.1, 0.0]),
-        rtol=1e-5,
-        atol=1e-6,
-    )
-
-
-def test_ltx25_checkpoint_selects_distilled_two_stage_profile(tmp_path, monkeypatch):
-    from vllm_omni.diffusion.models.ltx2 import ltx2_runtime
-
-    (tmp_path / "model_index.json").write_text(
-        json.dumps({"text_encoder": ["transformers", "Gemma4UnifiedForConditionalGeneration"]})
-    )
-
-    def stub_components(pipe, od_config):
-        pipe.od_config = od_config
-        pipe.vae_spatial_compression_ratio = 32
-
-    monkeypatch.setattr(ltx2_runtime, "initialize_pipeline_components", stub_components)
-    monkeypatch.setattr(LTXRuntime, "setup_diffusion_pipeline_profiler", lambda *_args, **_kwargs: None)
-
-    pipe = LTX2DistilledTwoStagePipeline(
-        od_config=SimpleNamespace(
-            model=str(tmp_path),
-            enable_diffusion_pipeline_profiler=False,
-        )
-    )
-
-    assert pipe.component_profile is LTX25_DISTILLED_COMPONENT_PROFILE
-    assert pipe.pipeline_recipe is LTX25_DISTILLED_TWO_STAGE_RECIPE
-
-
-def test_ltx25_full_config_enrichment_uses_selected_transformer_subfolder(tmp_path):
-    from vllm_omni.diffusion.data import OmniDiffusionConfig
-
-    (tmp_path / "model_index.json").write_text(json.dumps({"_class_name": "LTX2Pipeline", "model_version": "2.5"}))
-    for subfolder, marker in (("transformer", "distilled"), ("transformer_full", "full")):
-        path = tmp_path / subfolder
-        path.mkdir()
-        (path / "config.json").write_text(json.dumps({"marker": marker}))
-
-    config = OmniDiffusionConfig(model=str(tmp_path), model_class_name="LTX2Pipeline")
-    config.enrich_config()
-
-    assert config.tf_model_config.get("marker") == "full"
-
-
-def test_ltx25_four_public_pipeline_semantics_are_disjoint():
-    assert resolve_ltx_component_profile("one_stage", "2.5") is LTX25_FULL_COMPONENT_PROFILE
-    assert resolve_ltx_pipeline_recipe("one_stage", "2.5") is LTX25_FULL_RECIPE
-    assert resolve_ltx_component_profile("two_stage", "2.5") is LTX25_TWO_STAGE_COMPONENT_PROFILE
-    assert resolve_ltx_pipeline_recipe("two_stage", "2.5") is LTX25_TWO_STAGE_RECIPE
-    assert resolve_ltx_component_profile("distilled_one_stage", "2.5") is LTX25_DISTILLED_ONE_STAGE_COMPONENT_PROFILE
-    assert resolve_ltx_pipeline_recipe("distilled_one_stage", "2.5") is LTX25_DISTILLED_ONE_STAGE_RECIPE
-    assert resolve_ltx_component_profile("distilled_two_stage", "2.5") is LTX25_DISTILLED_COMPONENT_PROFILE
-    assert resolve_ltx_pipeline_recipe("distilled_two_stage", "2.5") is LTX25_DISTILLED_TWO_STAGE_RECIPE
-
-    assert LTX2Pipeline.pipeline_kind == "one_stage"
-    assert LTX2TwoStagePipeline.pipeline_kind == "two_stage"
-    assert LTX2DistilledOneStagePipeline.pipeline_kind == "distilled_one_stage"
-    assert LTX2DistilledTwoStagePipeline.pipeline_kind == "distilled_two_stage"
-
-
-@pytest.mark.parametrize(
-    ("pipeline_kind", "expected"),
-    [
-        ("one_stage", "regular"),
-        ("two_stage", "regular"),
-        ("distilled_one_stage", "distilled"),
-        ("distilled_two_stage", "distilled"),
-        ("dmd2", None),
-    ],
-)
-def test_ltx_checkpoint_kind_is_derived_from_pipeline_kind(pipeline_kind, expected):
-    assert resolve_ltx_checkpoint_kind(pipeline_kind) == expected
-
-
-def test_ltx_checkpoint_kind_rejects_unknown_pipeline():
-    with pytest.raises(ValueError, match="Unsupported LTX pipeline kind"):
-        resolve_ltx_checkpoint_kind("unknown")
-
-
-def test_ltx25_distilled_two_stage_recipe_matches_model_card_resolution():
-    stage1, stage2 = LTX25_DISTILLED_TWO_STAGE_RECIPE.phases
-
-    assert (LTX25_DISTILLED_TWO_STAGE_RECIPE.height, LTX25_DISTILLED_TWO_STAGE_RECIPE.width) == (1088, 1920)
-    assert stage1.spatial_downscale == 2
-    assert stage1.sigmas == LTX_DISTILLED_SIGMAS
-    assert stage1.sampler == "euler_ancestral"
-    assert stage2.sampler == "euler"
-    assert stage2.sigmas == (0.909375, 0.725, 0.421875, 0.0)
-    assert stage2.input_transform == "spatial_upsample"
-
-    assert LTX25_DISTILLED_TWO_STAGE_RECIPE.negative_prompt == ""
-    assert not LTX25_DISTILLED_TWO_STAGE_RECIPE.allow_negative_prompt
-
-
-def test_ltx25_distilled_one_stage_uses_native_low_resolution():
-    (phase,) = LTX25_DISTILLED_ONE_STAGE_RECIPE.phases
-
-    assert (LTX25_DISTILLED_ONE_STAGE_RECIPE.height, LTX25_DISTILLED_ONE_STAGE_RECIPE.width) == (544, 960)
-    assert phase.spatial_downscale == 1
-    assert phase.sigmas == LTX_DISTILLED_SIGMAS
-    assert phase.sampler == "euler_ancestral"
-    assert not LTX25_DISTILLED_ONE_STAGE_RECIPE.supports_cache_dit
-
-
-def test_ltx25_regular_two_stage_uses_full_transformer_then_distilled_lora():
-    stage1, stage2 = LTX25_TWO_STAGE_RECIPE.phases
-
-    assert (LTX25_TWO_STAGE_RECIPE.height, LTX25_TWO_STAGE_RECIPE.width) == (1088, 1920)
-    assert LTX25_TWO_STAGE_RECIPE.num_inference_steps == 30
-    assert stage1.spatial_downscale == 2
-    assert stage1.guidance == LTX25_FULL_RECIPE.request_guidance
-    assert stage2.adapter_slot == "ltx_distilled"
-    assert LTX25_TWO_STAGE_RECIPE.allow_request_phase_sigmas
-    assert LTX25_TWO_STAGE_COMPONENT_PROFILE.transformer_subfolder == "transformer_full"
-    assert LTX25_TWO_STAGE_COMPONENT_PROFILE.distilled_lora_filename is not None
-
-
-@pytest.mark.parametrize(
-    "recipe",
-    [
-        LTX2_ONE_STAGE_RECIPE,
-        LTX23_ONE_STAGE_RECIPE,
-        LTX_POSITIVE_ONLY_RECIPE,
-    ],
-)
-def test_ltx_one_stage_recipes_declare_cache_dit_supported(recipe):
-    assert recipe.supports_cache_dit
-
-
-@pytest.mark.parametrize(
-    "recipe",
-    [
-        LTX25_FULL_RECIPE,
-        LTX25_DISTILLED_ONE_STAGE_RECIPE,
-        LTX2_DISTILLED_TWO_STAGE_RECIPE,
-        LTX25_DISTILLED_TWO_STAGE_RECIPE,
-    ],
-)
-def test_ltx25_and_multistage_recipes_declare_cache_dit_unsupported(recipe):
-    assert not recipe.supports_cache_dit
-
-
-def test_ltx_positive_only_guidance_preserves_official_x0_roundtrip():
-    sample = torch.full((1, 1), 1e8, dtype=torch.float32)
-    velocity = torch.ones((1, 1), dtype=torch.bfloat16)
-    sigma = torch.tensor(0.5)
-
-    actual = LTX_GUIDANCE_EXECUTOR._guide_modality(
-        sample,
-        {"cond": velocity},
-        sigma,
-        LTXModalityGuidance(),
-    )
-    expected = velocity_from_x0(sample, x0_from_velocity(sample, velocity, sigma), sigma)
-
-    assert actual.dtype == torch.float32
-    torch.testing.assert_close(actual, expected, rtol=0.0, atol=0.0)
-
-
-def test_ltx_velocity_from_x0_materializes_official_scalar_sigma():
-    item_calls = []
-
-    class TrackingTensor(torch.Tensor):
-        @staticmethod
-        def __new__(cls):
-            return torch.Tensor._make_subclass(cls, torch.tensor(0.725), False)
-
-        def item(self):
-            item_calls.append(True)
-            return super().item()
-
-    sample = torch.tensor([[1.0]], dtype=torch.bfloat16)
-    x0 = torch.tensor([[0.5]], dtype=torch.bfloat16)
-
-    velocity_from_x0(sample, x0, TrackingTensor())
-
-    assert item_calls == [True]
-
-
-def test_ltx_ancestral_positive_only_guidance_preserves_raw_velocity():
-    sample = torch.full((1, 1), 1e8, dtype=torch.float32)
-    velocity = torch.ones((1, 1), dtype=torch.bfloat16)
-
-    actual = LTX_GUIDANCE_EXECUTOR._guide_modality(
-        sample,
-        {"cond": velocity},
-        torch.tensor(0.5),
-        LTXModalityGuidance(),
-        preserve_positive_velocity=True,
-    )
-
-    assert actual is velocity
-
-
-@pytest.mark.parametrize("model_version", ["2.5", "2.3"])
-def test_ltx_t2v_denoise_state_uses_official_bfloat16_dtype(model_version):
-    expected_dtype = torch.bfloat16
-    captured = {}
-
-    def prepare_latents(**kwargs):
-        captured["video_dtype"] = kwargs["dtype"]
-        return torch.empty(1, 1, 1, dtype=kwargs["dtype"])
-
-    def prepare_audio_latents(*_args, **kwargs):
-        captured["audio_dtype"] = kwargs["dtype"]
-        return torch.empty(1, 1, 1, dtype=kwargs["dtype"]), 1, 1
-
-    pipeline = SimpleNamespace(
-        model_version=model_version,
-        transformer=SimpleNamespace(config=SimpleNamespace(in_channels=128)),
-        prepare_latents=prepare_latents,
-        prepare_audio_latents=prepare_audio_latents,
-        audio_sampling_rate=48_000,
-        audio_hop_length=160,
-        audio_vae_temporal_compression_ratio=4,
-        audio_vae_mel_compression_ratio=4,
-        audio_vae=SimpleNamespace(config=SimpleNamespace(mel_bins=64, latent_channels=8)),
-        _resolve_audio_latent_length=lambda length, _latents: length,
-    )
-    request = SimpleNamespace(
-        num_videos_per_prompt=1,
-        height=64,
-        width=64,
-        num_frames=1,
-        frame_rate=24.0,
-        generator=None,
-        latents=None,
-        audio_latents=None,
-        audio_latents_normalized=False,
-    )
-    prompt = SimpleNamespace(
-        batch_size=1,
-        positive_connector_prompt_embeds=torch.empty(1, dtype=torch.bfloat16),
-        positive_connector_audio_prompt_embeds=torch.empty(1, dtype=torch.bfloat16),
-    )
-
-    LTXRuntime._prepare_video_latents_stage(
-        pipeline,
-        request,
-        prompt,
-        device=torch.device("cpu"),
-        noise_scale=1.0,
-    )
-    LTXRuntime._prepare_audio_latents_stage(
-        pipeline,
-        request,
-        prompt,
-        device=torch.device("cpu"),
-        noise_scale=1.0,
-    )
-
-    assert captured == {"video_dtype": expected_dtype, "audio_dtype": expected_dtype}
-
-
-def test_ltx_seeded_latents_match_official_packed_rng_layout():
-    pipeline = SimpleNamespace(
-        model_version="2.5",
-        vae_spatial_compression_ratio=2,
-        vae_temporal_compression_ratio=2,
-        transformer_spatial_patch_size=1,
-        transformer_temporal_patch_size=1,
-        audio_vae_mel_compression_ratio=2,
-        od_config=SimpleNamespace(parallel_config=SimpleNamespace(sequence_parallel_size=1)),
-    )
-    device = torch.device("cpu")
-    actual_generator = torch.Generator(device=device).manual_seed(123)
-    expected_generator = torch.Generator(device=device).manual_seed(123)
-
-    actual_video = ltx2_latents.prepare_video_latents(
-        pipeline,
-        batch_size=1,
-        num_channels_latents=2,
-        height=4,
-        width=6,
-        num_frames=3,
-        dtype=torch.float32,
-        device=device,
-        generator=actual_generator,
-    )
-    actual_audio, original_length, padded_length = ltx2_latents.prepare_audio_latents(
-        pipeline,
-        batch_size=1,
-        num_channels_latents=2,
-        audio_latent_length=3,
-        num_mel_bins=4,
-        dtype=torch.float32,
-        device=device,
-        generator=actual_generator,
-    )
-
-    expected_video = torch.randn((1, 12, 2), generator=expected_generator)
-    expected_audio = torch.randn((1, 3, 4), generator=expected_generator)
-
-    torch.testing.assert_close(actual_video, expected_video, rtol=0.0, atol=0.0)
-    torch.testing.assert_close(actual_audio, expected_audio, rtol=0.0, atol=0.0)
-    assert (original_length, padded_length) == (3, 3)
-
-
-def test_ltx25_distilled_two_stage_executes_custom_phase_schedules():
-    request_inputs = LTXRequestInputs(
-        prompt="prompt",
-        negative_prompt="",
-        height=64,
-        width=64,
-        num_frames=1,
-        frame_rate=24.0,
-        num_inference_steps=8,
-        guidance=LTXGuidanceSpec.positive_only(),
-        num_videos_per_prompt=1,
-        generator=None,
-        latents=None,
-        audio_latents=None,
-        prompt_embeds=None,
-        negative_prompt_embeds=None,
-        prompt_attention_mask=None,
-        negative_prompt_attention_mask=None,
-        decode_timestep=0.25,
-        decode_noise_scale=0.5,
-        output_type="np",
-        max_sequence_length=16,
-    )
-    prompt_context = object()
-    phase_calls = []
-
-    def resolve_request_inputs(req, **kwargs):
-        return request_inputs
-
-    def run_phase(req, inputs, *, prompt_context=None, **kwargs):
-        phase_recipe = kwargs["phase_recipe"]
-        assert kwargs["image"] is source_image
-        phase_calls.append((phase_recipe, inputs, prompt_context))
-        if len(phase_calls) == 1:
-            assert prompt_context is None
-            assert inputs.decode_timestep == 0.25
-            assert inputs.decode_noise_scale == 0.5
-            context = prompt_context_sentinel
-            assert phase_recipe.name == "generate_lowres"
-            assert (inputs.height, inputs.width) == (32, 32)
-            assert inputs.num_inference_steps == 2
-            assert kwargs["sigmas"] == stage_1_sigmas
-            assert kwargs["noise_scale"] == 1.0
-            video = torch.ones(1, 128, 1, 1, 1)
-            audio = torch.full((1, 8, 1, 2), 2.0)
-            audio_for_next_phase = torch.full((1, 8, 1, 2), 5.0)
-        else:
-            assert prompt_context is prompt_context_sentinel
-            assert phase_recipe.name == "refine"
-            assert (inputs.height, inputs.width) == (64, 64)
-            assert inputs.latents.shape == (1, 128, 1, 2, 2)
-            torch.testing.assert_close(inputs.audio_latents, torch.full((1, 8, 1, 2), 5.0))
-            assert inputs.audio_latents_normalized
-            assert inputs.guidance_scale == 1.0
-            assert inputs.num_inference_steps == 2
-            assert kwargs["sigmas"] == stage_2_sigmas
-            assert kwargs["noise_scale"] == stage_2_sigmas[0]
-            assert inputs.decode_timestep == 0.0
-            assert inputs.decode_noise_scale is None
-            context = prompt_context
-            video = torch.full((1, 128, 1, 2, 2), 3.0)
-            audio = torch.full((1, 8, 1, 2), 4.0)
-            audio_for_next_phase = None
-        return LTXPhaseResult(
-            forward_context=SimpleNamespace(prompt_context=context),
-            video=video,
-            audio=audio,
-            audio_for_next_phase=audio_for_next_phase,
-        )
-
-    def decode_phase(phase):
-        return DiffusionOutput(output=(phase.video, phase.audio))
-
-    class FakeUpsampler(torch.nn.Module):
-        dtype = torch.float32
-
-        def forward(self, latents):
-            return latents.repeat_interleave(2, dim=-2).repeat_interleave(2, dim=-1)
-
-    prompt_context_sentinel = prompt_context
-    stage_1_sigmas = [1.0, 0.5, 0.0]
-    stage_2_sigmas = [0.8, 0.2, 0.0]
-    pipeline = object.__new__(LTX2TwoStagePipeline)
-    torch.nn.Module.__init__(pipeline)
-    pipeline.pipeline_recipe = LTX25_DISTILLED_TWO_STAGE_RECIPE
-    pipeline.device = torch.device("cpu")
-    pipeline.vae_spatial_compression_ratio = 32
-    pipeline.vae_temporal_compression_ratio = 8
-    pipeline.latent_upsampler = FakeUpsampler()
-    object.__setattr__(pipeline, "_resolve_request_inputs", resolve_request_inputs)
-    object.__setattr__(pipeline, "run_phase", run_phase)
-    object.__setattr__(pipeline, "decode_phase", decode_phase)
-
-    source_image = object()
-    output = pipeline.forward(
-        SimpleNamespace(sampling_params_list=[], prompts=[]),
-        image=source_image,
-        stage_1_sigmas=stage_1_sigmas,
-        stage_2_sigmas=stage_2_sigmas,
-    )
-
-    assert len(phase_calls) == 2
-    assert phase_calls[1][2] is prompt_context_sentinel
-    torch.testing.assert_close(output.output[0], torch.full((1, 128, 1, 2, 2), 3.0))
-    torch.testing.assert_close(output.output[1], torch.full((1, 8, 1, 2), 4.0))
-
-
-class TestLTXRequestParsing:
-    def test_request_resolves_custom_sigmas_from_extra_args(self):
-        from vllm_omni.diffusion.request import OmniDiffusionRequest
-        from vllm_omni.diffusion.worker.request_batch import DiffusionRequestBatch
-        from vllm_omni.inputs.data import OmniDiffusionSamplingParams
-
-        sigmas = [1.0, 0.5, 0.0]
-        req = DiffusionRequestBatch(
-            [
-                OmniDiffusionRequest(
-                    prompt="prompt",
-                    sampling_params=OmniDiffusionSamplingParams(extra_args={"sigmas": sigmas}),
-                    request_id="ltx-custom-sigmas",
-                )
-            ]
-        )
-
-        assert _make_ltx_request_pipe(LTX2Pipeline)._resolve_request_sigmas(req, None) == sigmas
-
-    def test_request_resolves_per_stage_sigmas_from_extra_args(self):
-        from vllm_omni.diffusion.request import OmniDiffusionRequest
-        from vllm_omni.diffusion.worker.request_batch import DiffusionRequestBatch
-        from vllm_omni.inputs.data import OmniDiffusionSamplingParams
-
-        stage_1_sigmas = [1.0, 0.4, 0.0]
-        req = DiffusionRequestBatch(
-            [
-                OmniDiffusionRequest(
-                    prompt="prompt",
-                    sampling_params=OmniDiffusionSamplingParams(extra_args={"stage_1_sigmas": stage_1_sigmas}),
-                    request_id="ltx-custom-phase-sigmas",
-                )
-            ]
-        )
-
-        resolved = _make_ltx_request_pipe(LTX2TwoStagePipeline)._resolve_request_phase_sigmas(
-            req, [1.0, 0.0], [0.8, 0.0]
-        )
-
-        assert resolved == (stage_1_sigmas, [0.8, 0.0])
-
-    @pytest.mark.parametrize("value", [-1, 52, 1.5, True])
-    def test_request_rejects_invalid_image_crf(self, value):
-        from vllm_omni.diffusion.request import OmniDiffusionRequest
-        from vllm_omni.diffusion.worker.request_batch import DiffusionRequestBatch
-        from vllm_omni.inputs.data import OmniDiffusionSamplingParams
-
-        req = DiffusionRequestBatch(
-            [
-                OmniDiffusionRequest(
-                    prompt="prompt",
-                    sampling_params=OmniDiffusionSamplingParams(extra_args={"image_crf": value}),
-                    request_id="ltx-invalid-image-crf",
-                )
-            ]
-        )
-
-        with pytest.raises(ValueError, match="image_crf"):
-            _resolve_request_inputs_for_test(_make_ltx_request_pipe(LTX2Pipeline), req)
-
-
-class TestLTXForwardStages:
-    def test_gemma_tokenization_matches_official_bos_and_left_padding(self):
-        pipe = _make_ltx_request_pipe(LTX2Pipeline)
-
-        class FakeTokenizer:
-            bos_token_id = 2
-            eos_token_id = 1
-            eos_token = "<eos>"
-            pad_token_id = 0
-            pad_token = "<pad>"
-            padding_side = "right"
-
-            def __init__(self):
-                self.calls = []
-
-            def __call__(self, text, **kwargs):
-                self.calls.append((text, kwargs))
-                ids = [7, 8, 9, self.eos_token_id] if text == "long" else [6, self.eos_token_id]
-                return SimpleNamespace(input_ids=ids)
-
-            def pad(self, encoded, *, padding, max_length, return_tensors, return_attention_mask):
-                assert padding == "max_length"
-                assert return_tensors == "pt"
-                assert return_attention_mask is True
-                rows = []
-                masks = []
-                for ids in encoded["input_ids"]:
-                    pad = max_length - len(ids)
-                    rows.append([self.pad_token_id] * pad + ids)
-                    masks.append([0] * pad + [1] * len(ids))
-                return SimpleNamespace(input_ids=torch.tensor(rows), attention_mask=torch.tensor(masks))
-
-        class FakeTextEncoder(torch.nn.Module):
-            dtype = torch.float32
-
-            def __init__(self):
-                super().__init__()
-                self.seen_input_ids = None
-                self.seen_attention_mask = None
-
-            def forward(self, *, input_ids, attention_mask, output_hidden_states):
-                assert output_hidden_states is True
-                self.seen_input_ids = input_ids.clone()
-                self.seen_attention_mask = attention_mask.clone()
-                return SimpleNamespace(hidden_states=(input_ids.unsqueeze(-1).float(),))
-
-        tokenizer = FakeTokenizer()
-        text_encoder = FakeTextEncoder()
-        object.__setattr__(pipe, "tokenizer", tokenizer)
-        object.__setattr__(pipe, "text_encoder", text_encoder)
-
-        pipe._get_gemma_prompt_embeds(
-            [" long ", "short"],
-            max_sequence_length=4,
-            device=torch.device("cpu"),
-            dtype=torch.float32,
-        )
-
-        torch.testing.assert_close(text_encoder.seen_input_ids, torch.tensor([[2, 7, 8, 9], [0, 2, 6, 1]]))
-        torch.testing.assert_close(text_encoder.seen_attention_mask, torch.tensor([[1, 1, 1, 1], [0, 1, 1, 1]]))
-        assert tokenizer.padding_side == "left"
-        assert [call[0] for call in tokenizer.calls] == ["long", "short"]
@@ -3,7 +3,6 @@

 """Unit tests for LTX image-to-video input and conditioning behavior."""

-import sys
 from types import SimpleNamespace

 import pytest
@@ -24,7 +23,7 @@ def _make_ltx23_request_pipe(cls):


 class TestLTXImageToVideoForwardStages:
-    def test_legacy_i2v_pil_preprocessing_preserves_aspect_ratio_and_center_crops(self):
+    def test_i2v_pil_preprocessing_matches_official_center_crop(self):
        from vllm_omni.diffusion.models.ltx2.ltx2_conditioning import _preprocess_i2v_pil_images

        pixels = torch.zeros(4, 8, 3, dtype=torch.uint8)
@@ -36,136 +35,6 @@ class TestLTXImageToVideoForwardStages:

        torch.testing.assert_close(actual, expected)

-    def test_ltx25_i2v_pil_preprocessing_preserves_aspect_ratio_after_crf(self, monkeypatch):
-        import vllm_omni.diffusion.models.ltx2.ltx2_conditioning as conditioning
-
-        pixels = torch.zeros(4, 8, 3, dtype=torch.uint8)
-        pixels[:, :, 0] = torch.arange(8, dtype=torch.uint8)
-        image = Image.fromarray(pixels.numpy())
-        monkeypatch.setattr(conditioning, "_apply_image_conditioning_crf", lambda image_array, _crf: image_array)
-
-        actual = conditioning._preprocess_i2v_pil_images(image, height=4, width=4, crf=18)
-        expected = pixels[:, 2:6].permute(2, 0, 1).unsqueeze(0).float() / 127.5 - 1.0
-
-        torch.testing.assert_close(actual, expected)
-
-    def test_ltx25_i2v_applies_crf18_before_resize(self, monkeypatch):
-        import vllm_omni.diffusion.models.ltx2.ltx2_conditioning as conditioning
-        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline
-
-        captured_crfs = []
-        captured_prepare_kwargs = []
-
-        def fake_preprocess(image, *, height, width, crf, device, dtype):
-            del image, device
-            captured_crfs.append(crf)
-            return torch.zeros(1, 3, height, width, dtype=dtype)
-
-        def fake_prepare_latents(**kwargs):
-            captured_prepare_kwargs.append(kwargs)
-            return kwargs["image"], None
-
-        monkeypatch.setattr(conditioning, "_preprocess_i2v_pil_images", fake_preprocess)
-        pipe = object.__new__(LTX2Pipeline)
-        torch.nn.Module.__init__(pipe)
-        pipe.model_version = "2.5"
-        object.__setattr__(pipe, "transformer", SimpleNamespace(config=SimpleNamespace(in_channels=4)))
-        object.__setattr__(pipe, "prepare_latents", fake_prepare_latents)
-        request_inputs = SimpleNamespace(
-            height=4,
-            width=4,
-            num_frames=3,
-            num_videos_per_prompt=1,
-            generator=None,
-            latents=None,
-        )
-        prompt_context = SimpleNamespace(
-            batch_size=1,
-            positive_connector_prompt_embeds=torch.zeros(1, 1, 1, dtype=torch.bfloat16),
-        )
-
-        pipe._prepare_video_latents_stage(
-            request_inputs,
-            prompt_context,
-            device=torch.device("cpu"),
-            noise_scale=0.0,
-            image=Image.new("RGB", (8, 4)),
-        )
-
-        assert captured_crfs == [18]
-        assert captured_prepare_kwargs[0]["dtype"] is torch.bfloat16
-        assert captured_prepare_kwargs[0]["image"].dtype is torch.bfloat16
-
-    def test_ltx25_i2v_rejects_inputs_that_cannot_receive_crf18(self):
-        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline
-
-        pipe = object.__new__(LTX2Pipeline)
-        torch.nn.Module.__init__(pipe)
-        pipe.model_version = "2.5"
-
-        with pytest.raises(ValueError, match="requires PIL images"):
-            pipe._prepare_video_latents_stage(
-                SimpleNamespace(),
-                SimpleNamespace(),
-                device=torch.device("cpu"),
-                noise_scale=0.0,
-                image=torch.zeros(3, 4, 4),
-            )
-
-    def test_ltx25_i2v_tensor_input_can_opt_out_of_crf18(self):
-        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline
-
-        captured_prepare_kwargs = []
-
-        def fake_prepare_latents(**kwargs):
-            captured_prepare_kwargs.append(kwargs)
-            return kwargs["image"], None
-
-        pipe = object.__new__(LTX2Pipeline)
-        torch.nn.Module.__init__(pipe)
-        pipe.model_version = "2.5"
-        object.__setattr__(pipe, "transformer", SimpleNamespace(config=SimpleNamespace(in_channels=4)))
-        object.__setattr__(pipe, "prepare_latents", fake_prepare_latents)
-        request_inputs = SimpleNamespace(
-            height=4,
-            width=4,
-            num_frames=3,
-            num_videos_per_prompt=1,
-            generator=None,
-            latents=None,
-            image_crf=0,
-        )
-        prompt_context = SimpleNamespace(
-            batch_size=1,
-            positive_connector_prompt_embeds=torch.zeros(1, 1, 1, dtype=torch.bfloat16),
-        )
-
-        pipe._prepare_video_latents_stage(
-            request_inputs,
-            prompt_context,
-            device=torch.device("cpu"),
-            noise_scale=0.0,
-            image=torch.zeros(3, 4, 4),
-        )
-
-        assert captured_prepare_kwargs[0]["image"].shape == (1, 3, 4, 4)
-
-    def test_ltx25_i2v_crf_preserves_single_pixel_dimension(self, monkeypatch):
-        from vllm_omni.diffusion.models.ltx2.ltx2_conditioning import _apply_image_conditioning_crf
-
-        image = torch.zeros(1, 4, 3, dtype=torch.uint8).numpy()
-        monkeypatch.setitem(sys.modules, "av", None)
-
-        assert _apply_image_conditioning_crf(image, 18) is image
-
-    def test_ltx25_i2v_reports_missing_pyav(self, monkeypatch):
-        from vllm_omni.diffusion.models.ltx2.ltx2_conditioning import _apply_image_conditioning_crf
-
-        monkeypatch.setitem(sys.modules, "av", None)
-
-        with pytest.raises(ImportError, match="PyAV with a libx264 encoder"):
-            _apply_image_conditioning_crf(torch.zeros(4, 4, 3, dtype=torch.uint8).numpy(), 18)
-
    def test_forward_resolves_request_image_and_delegates_to_shared_recipe_runtime(self):
        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline
        from vllm_omni.diffusion.request import OmniDiffusionRequest
@@ -229,7 +98,7 @@ class TestLTXImageToVideoForwardStages:
        with pytest.raises(ValueError, match="cannot mix text-to-video and image-to-video"):
            pipe._resolve_request_image(mixed_req, None, request_inputs)

-    def test_denoise_timestep_kwargs_masks_video_and_keeps_audio_per_token(self):
+    def test_denoise_timestep_kwargs_masks_video_only(self):
        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline

        pipe = object.__new__(LTX2Pipeline)
@@ -320,18 +189,13 @@ class TestLTXImageToVideoConditioning:
        torch.testing.assert_close(conditioning_mask, torch.tensor([[1.0, 0.0, 0.0]]))
        torch.testing.assert_close(out, torch.tensor([[[40.0, 41.0], [20.0, 21.0], [30.0, 31.0]]]))

-    @pytest.mark.parametrize(
-        ("model_version", "sampled_shape"),
-        [("2.3", (1, 3, 2)), ("2.5", (1, 3, 2))],
-    )
-    def test_i2v_5d_latents_noise_uses_packed_rng_layout(self, monkeypatch, model_version, sampled_shape):
+    def test_ltx23_i2v_5d_latents_noise_preserves_conditioning_frame(self, monkeypatch):
        import vllm_omni.diffusion.models.ltx2.ltx2_conditioning as ltx2_conditioning
        import vllm_omni.diffusion.models.ltx2.ltx2_latents as ltx2_latents
        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline

        pipe = object.__new__(LTX2Pipeline)
        torch.nn.Module.__init__(pipe)
-        pipe.model_version = model_version
        pipe.vae_spatial_compression_ratio = 1
        pipe.vae_temporal_compression_ratio = 1
        pipe.transformer_spatial_patch_size = 1
@@ -373,20 +237,14 @@ class TestLTXImageToVideoConditioning:

        torch.testing.assert_close(conditioning_mask, torch.tensor([[1.0, 0.0, 0.0]]))
        torch.testing.assert_close(out, torch.tensor([[[40.0, 41.0], [1.0, 1.0], [1.0, 1.0]]]))
-        assert sampled_shapes == [sampled_shape]
+        assert sampled_shapes == [(1, 3, 2)]

-    @pytest.mark.parametrize(
-        ("model_version", "sampled_shape"),
-        [("2.3", (1, 3, 2)), ("2.5", (1, 3, 2))],
-    )
-    def test_i2v_image_noise_uses_packed_rng_layout(self, monkeypatch, model_version, sampled_shape):
+    def test_ltx23_i2v_image_noise_is_sampled_after_packing(self, monkeypatch):
        import vllm_omni.diffusion.models.ltx2.ltx2_conditioning as ltx2_conditioning
-        import vllm_omni.diffusion.models.ltx2.ltx2_latents as ltx2_latents
        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline

        pipe = object.__new__(LTX2Pipeline)
        torch.nn.Module.__init__(pipe)
-        pipe.model_version = model_version
        pipe.vae_spatial_compression_ratio = 1
        pipe.vae_temporal_compression_ratio = 1
        pipe.transformer_spatial_patch_size = 1
@@ -408,7 +266,6 @@ class TestLTXImageToVideoConditioning:
            return torch.ones(shape, device=device, dtype=dtype)

        monkeypatch.setattr(ltx2_conditioning, "randn_tensor", fake_randn_tensor)
-        monkeypatch.setattr(ltx2_latents, "randn_tensor", fake_randn_tensor)

        out, conditioning_mask = pipe.prepare_latents(
            image=torch.zeros(1, 3, 1, 1),
@@ -422,54 +279,10 @@ class TestLTXImageToVideoConditioning:
            device=torch.device("cpu"),
        )

-        assert sampled_shapes == [sampled_shape]
+        assert sampled_shapes == [(1, 3, 2)]
        torch.testing.assert_close(conditioning_mask, torch.tensor([[1.0, 0.0, 0.0]]))
        torch.testing.assert_close(out, torch.tensor([[[10.0, 11.0], [1.0, 1.0], [1.0, 1.0]]]))

-    def test_ltx25_re_noise_uses_official_scalar_noise_scale(self, monkeypatch):
-        import vllm_omni.diffusion.models.ltx2.ltx2_latents as ltx2_latents
-        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline
-
-        pipe = object.__new__(LTX2Pipeline)
-        torch.nn.Module.__init__(pipe)
-        pipe.model_version = "2.5"
-        pipe.vae_spatial_compression_ratio = 1
-        pipe.vae_temporal_compression_ratio = 1
-        pipe.transformer_spatial_patch_size = 1
-        pipe.transformer_temporal_patch_size = 1
-        pipe.vae = SimpleNamespace(
-            latents_mean=torch.zeros(2),
-            latents_std=torch.ones(2),
-            config=SimpleNamespace(scaling_factor=1.0),
-        )
-        object.__setattr__(
-            pipe,
-            "_encode_i2v_image_latents",
-            lambda *_args, **_kwargs: torch.zeros(1, 2, 1, 1, 1, dtype=torch.bfloat16),
-        )
-        captured = {}
-
-        def fake_create_noised_state(latents, noise_scale, generator=None):
-            del generator
-            captured["noise_scale"] = noise_scale
-            return latents
-
-        monkeypatch.setattr(ltx2_latents, "create_noised_state", fake_create_noised_state)
-        pipe.prepare_latents(
-            image=torch.zeros(1, 3, 1, 1),
-            batch_size=1,
-            num_channels_latents=2,
-            height=1,
-            width=1,
-            num_frames=3,
-            noise_scale=0.909375,
-            dtype=torch.bfloat16,
-            device=torch.device("cpu"),
-            latents=torch.zeros(1, 2, 3, 1, 1, dtype=torch.bfloat16),
-        )
-
-        assert captured["noise_scale"] == 0.909375
-
    def test_ltx23_i2v_video_step_preserves_conditioning_frame(self):
        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline

@@ -572,99 +385,6 @@ class TestLTXImageToVideoConditioning:
        torch.testing.assert_close(audio, torch.tensor([[[-2.0]]]))
        assert pipeline.step_indices == [0, 1]

-    def test_ltx25_ancestral_adapter_matches_official_seeded_step_and_preserves_i2v_mask(self):
-        from vllm_omni.diffusion.models.ltx2.ltx2_denoise import LTXVideoAudioStepAdapter
-
-        device = torch.device("cpu")
-        sigmas = torch.tensor([1.0, 0.5, 0.0])
-        pipeline = SimpleNamespace(device=device, scheduler=SimpleNamespace(sigmas=sigmas))
-        mask = torch.tensor([[1.0, 0.0]])
-        adapter = LTXVideoAudioStepAdapter(
-            pipeline,
-            SimpleNamespace(sigmas=sigmas.clone()),
-            latent_num_frames=1,
-            latent_height=1,
-            latent_width=1,
-            image_conditioned=True,
-            sampler="euler_ancestral",
-            generator=torch.Generator(device=device).manual_seed(42),
-            conditioning_mask=mask,
-        )
-        video = torch.tensor([[[3.0], [2.0]]])
-        audio = torch.tensor([[[4.0]]])
-        video_velocity = torch.tensor([[[7.0], [0.25]]])
-        audio_velocity = torch.tensor([[[0.5]]])
-
-        reference_generator = torch.Generator(device=device).manual_seed(10042)
-        video_noise = torch.randn(video.shape, generator=reference_generator)
-        audio_noise = torch.randn(audio.shape, generator=reference_generator)
-
-        def official_step(sample, velocity, noise):
-            sigma = sigmas[0]
-            sigma_next = sigmas[1]
-            denoised = sample.float() - velocity.float() * sigma
-            sigma_down = sigma_next * (sigma_next / sigma)
-            ratio = sigma_down / sigma
-            deterministic = ratio * sample.float() + (1.0 - ratio) * denoised
-            alpha_next = 1.0 - sigma_next
-            alpha_down = 1.0 - sigma_down
-            coeff = (sigma_next**2 - sigma_down**2 * alpha_next**2 / alpha_down**2).sqrt()
-            return (alpha_next / alpha_down) * deterministic + noise * coeff
-
-        expected_video = official_step(video, video_velocity, video_noise)
-        expected_video[:, :1] = video[:, :1]
-        expected_audio = official_step(audio, audio_velocity, audio_noise)
-
-        ((actual_video, actual_audio),) = adapter.step(
-            (video_velocity, audio_velocity),
-            (sigmas[0], sigmas[0]),
-            (video, audio),
-        )
-        torch.testing.assert_close(actual_video, expected_video)
-        torch.testing.assert_close(actual_audio, expected_audio)
-
-        terminal_video_velocity = torch.full_like(actual_video, 0.125)
-        terminal_audio_velocity = torch.full_like(actual_audio, 0.25)
-        ((terminal_video, terminal_audio),) = adapter.step(
-            (terminal_video_velocity, terminal_audio_velocity),
-            (sigmas[1], sigmas[1]),
-            (actual_video, actual_audio),
-        )
-        expected_terminal_video = actual_video - terminal_video_velocity * sigmas[1]
-        expected_terminal_video[:, :1] = actual_video[:, :1]
-        expected_terminal_audio = actual_audio - terminal_audio_velocity * sigmas[1]
-        torch.testing.assert_close(terminal_video, expected_terminal_video)
-        torch.testing.assert_close(terminal_audio, expected_terminal_audio)
-
-    def test_ltx25_ancestral_quantizes_x0_to_model_dtype_before_fp32_step(self):
-        from vllm_omni.diffusion.models.ltx2.ltx2_denoise import _ancestral_euler_step_from_velocity
-
-        sample = torch.tensor([[[3.140625, -1.234375]]], dtype=torch.bfloat16)
-        velocity = torch.tensor([[[0.333984375, 2.71875]]], dtype=torch.bfloat16)
-        sigmas = torch.tensor([0.73, 0.41, 0.0], dtype=torch.float32)
-        generator = torch.Generator(device="cpu").manual_seed(10042)
-        reference_generator = torch.Generator(device="cpu").manual_seed(10042)
-
-        sigma, sigma_next = sigmas[:2]
-        denoised = (sample.float() - velocity.float() * sigma).to(sample.dtype).float()
-        sigma_down = sigma_next * (sigma_next / sigma)
-        sigma_down_ratio = sigma_down / sigma
-        expected = sigma_down_ratio * sample.float() + (1.0 - sigma_down_ratio) * denoised
-        alpha_next = 1.0 - sigma_next
-        alpha_down = 1.0 - sigma_down
-        coefficient = (sigma_next**2 - sigma_down**2 * alpha_next**2 / alpha_down**2).sqrt()
-        noise = torch.randn(
-            sample.shape,
-            generator=reference_generator,
-            dtype=sample.dtype,
-            device=sample.device,
-        )
-        expected = ((alpha_next / alpha_down) * expected + noise.float() * coefficient).to(sample.dtype)
-
-        actual = _ancestral_euler_step_from_velocity(sample, velocity, sigmas, 0, generator)
-
-        torch.testing.assert_close(actual, expected, rtol=0, atol=0)
-
    def test_i2v_guidance_uses_zero_sigma_for_conditioned_tokens(self):
        from vllm_omni.diffusion.models.ltx2.ltx2_denoise import LTXDenoiseContext
        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline
@@ -202,68 +202,3 @@ def test_prepare_audio_latents_rejects_incompatible_provided_length():
            device=torch.device("cpu"),
            latents=latents,
        )
-
-
-def test_create_noised_state_matches_official_fp32_lerp():
-    latents = torch.linspace(-2, 2, 4096, dtype=torch.bfloat16).reshape(1, 32, 128)
-    expected_generator = torch.Generator().manual_seed(42)
-    noise = torch.randn(latents.shape, generator=expected_generator, dtype=latents.dtype)
-    expected = torch.lerp(latents.float(), noise.float(), torch.tensor(0.15, dtype=torch.bfloat16).float()).to(
-        latents.dtype
-    )
-
-    actual = latent_ops.create_noised_state(
-        latents,
-        torch.tensor(0.15, dtype=torch.bfloat16),
-        torch.Generator().manual_seed(42),
-    )
-
-    torch.testing.assert_close(actual, expected, rtol=0, atol=0)
-
-
-def test_create_conditioned_noised_state_matches_official_two_lerps():
-    latents = torch.linspace(-2, 2, 24, dtype=torch.bfloat16).reshape(1, 3, 8)
-    clean_latents = latents.clone()
-    clean_latents[:, 0] = 3.0
-    denoise_mask = torch.tensor([[[0.0], [1.0], [1.0]]])
-    expected_generator = torch.Generator().manual_seed(42)
-    noise = torch.randn(latents.shape, generator=expected_generator, dtype=latents.dtype)
-    noised = torch.lerp(latents.float(), noise.float(), 0.909375)
-    expected = torch.lerp(clean_latents.float(), noised, denoise_mask).to(latents.dtype)
-
-    actual = latent_ops.create_conditioned_noised_state(
-        latents,
-        clean_latents,
-        denoise_mask,
-        0.909375,
-        torch.Generator().manual_seed(42),
-    )
-
-    torch.testing.assert_close(actual, expected, rtol=0, atol=0)
-
-
-def test_prepare_supplied_video_latents_uses_official_token_major_layout():
-    pipeline = _make_pipeline(LTX2Pipeline)
-    supplied = torch.arange(512, dtype=torch.float32).reshape(1, 32, 16)
-    actual_video = pipeline.prepare_latents(
-        batch_size=1,
-        num_channels_latents=4,
-        height=64,
-        width=64,
-        num_frames=9,
-        dtype=torch.float32,
-        device=torch.device("cpu"),
-        latents=supplied,
-    )
-
-    torch.testing.assert_close(actual_video, supplied, rtol=0, atol=0)
-    assert actual_video.stride()[1:] == (1, actual_video.shape[1])
-
-
-def test_clear_audio_padding_keeps_padding_outside_sampler_state():
-    updated = torch.tensor([[[1.0, 2.0], [3.0, 4.0], [99.0, -99.0]]])
-
-    actual = latent_ops.clear_audio_padding(updated, 2)
-
-    torch.testing.assert_close(actual[:, :2], updated[:, :2])
-    torch.testing.assert_close(actual[:, 2:], torch.zeros_like(actual[:, 2:]))
@@ -28,15 +28,6 @@ def test_velocity_from_x0_uses_official_host_scalar_sigma(mocker):
 class TestCFGParallelHelpers:
    """Test LTX-2.3 CFG helper math without loading model weights."""

-    def test_single_guidance_pass_preserves_official_token_layout(self):
-        from vllm_omni.diffusion.models.ltx2.ltx2_guidance import _repeat_batch
-
-        tokens = torch.randn(1, 16, 32).transpose(1, 2).contiguous().transpose(1, 2)
-        repeated = _repeat_batch(tokens, 1)
-
-        assert repeated is tokens
-        assert repeated.stride() == tokens.stride()
-
    def test_combine_cfg_noise_matches_x0_space_formula(self):
        from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import LTX2Pipeline

@@ -20,7 +20,6 @@ from vllm_omni.diffusion.models.ltx2 import (
 from vllm_omni.diffusion.models.ltx2.ltx2_components import (
    LTX2_COMPONENT_PROFILE,
    LTX2_DISTILLED_COMPONENT_PROFILE,
-    LTX2_DISTILLED_ONE_STAGE_COMPONENT_PROFILE,
    LTX2_TWO_STAGE_COMPONENT_PROFILE,
    LTX23_COMPONENT_PROFILE,
    LTX23_DISTILLED_COMPONENT_PROFILE,
@@ -48,7 +47,6 @@ from vllm_omni.diffusion.models.ltx2.ltx2_guidance import (
 )
 from vllm_omni.diffusion.models.ltx2.ltx2_latents import LTXAVState
 from vllm_omni.diffusion.models.ltx2.ltx2_recipes import (
-    LTX2_DISTILLED_ONE_STAGE_RECIPE,
    LTX2_DISTILLED_TWO_STAGE_RECIPE,
    LTX2_ONE_STAGE_RECIPE,
    LTX2_TWO_STAGE_RECIPE,
@@ -66,14 +64,12 @@ from vllm_omni.diffusion.models.ltx2.ltx2_request import (
 )
 from vllm_omni.diffusion.models.ltx2.ltx2_runtime import LTXRuntime
 from vllm_omni.diffusion.models.ltx2.pipeline_ltx2 import (
-    LTX2DistilledOneStagePipeline,
    LTX2I2VDMD2Pipeline,
    LTX2Pipeline,
    LTX2T2VDMD2Pipeline,
 )
 from vllm_omni.diffusion.models.ltx2.pipeline_ltx2_two_stage import (
    LTX2DistilledPipeline,
-    LTX2DistilledTwoStagePipeline,
    LTX2TwoStagePipeline,
 )

@@ -129,21 +125,15 @@ def test_ltx_public_entries_share_runtime_and_keep_recipe_boundaries():

    assert issubclass(LTX2Pipeline, LTXRuntime)
    assert issubclass(LTX2TwoStagePipeline, LTXRuntime)
-    assert issubclass(LTX2DistilledOneStagePipeline, LTXRuntime)
-    assert issubclass(LTX2DistilledTwoStagePipeline, LTXRuntime)
    assert issubclass(LTX2DistilledPipeline, LTXRuntime)
-    assert LTX2DistilledPipeline is LTX2DistilledTwoStagePipeline
    assert LTX2Pipeline.component_profile is LTX2_COMPONENT_PROFILE
    assert LTX2Pipeline.pipeline_recipe is LTX2_ONE_STAGE_RECIPE
    assert LTX2TwoStagePipeline.component_profile is LTX2_TWO_STAGE_COMPONENT_PROFILE
    assert LTX2TwoStagePipeline.pipeline_recipe is LTX2_TWO_STAGE_RECIPE
-    assert LTX2DistilledOneStagePipeline.component_profile is LTX2_DISTILLED_ONE_STAGE_COMPONENT_PROFILE
-    assert LTX2DistilledOneStagePipeline.pipeline_recipe is LTX2_DISTILLED_ONE_STAGE_RECIPE
    assert LTX2DistilledPipeline.component_profile is LTX2_DISTILLED_COMPONENT_PROFILE
    for pipeline_cls, profile in (
        (LTX2Pipeline, LTX2_COMPONENT_PROFILE),
        (LTX2TwoStagePipeline, LTX2_TWO_STAGE_COMPONENT_PROFILE),
-        (LTX2DistilledOneStagePipeline, LTX2_DISTILLED_ONE_STAGE_COMPONENT_PROFILE),
        (LTX2DistilledPipeline, LTX2_DISTILLED_COMPONENT_PROFILE),
    ):
        assert pipeline_cls._dit_modules == list(profile.dit_modules)
@@ -264,7 +254,7 @@ def test_ltx_multistage_rejects_cache_dit_before_component_initialization(tmp_pa

    monkeypatch.setattr(ltx2_runtime, "initialize_pipeline_components", stub_components)

-    with pytest.raises(ValueError, match="not qualified for this LTX recipe"):
+    with pytest.raises(ValueError, match="only one-stage LTX recipes"):
        LTX2TwoStagePipeline(
            od_config=SimpleNamespace(
                model=str(tmp_path),
@@ -719,48 +709,15 @@ def test_ltx_component_loader_installs_omni_connector_attention(monkeypatch):

    assert len(installed) == 2
    assert all(processor.__class__.__name__ == "_LTXConnectorAttnProcessor" for processor in installed)
+    assert all(processor.has_learned_registers for processor in installed)
    assert all(kernel["role"] == "ltx2.connector" for kernel in kernels)
    assert all(kernel["role_category"] == "self" for kernel in kernels)
    assert all(kernel["skip_sequence_parallel"] for kernel in kernels)
    assert all(kernel["disable_kv_quant"] for kernel in kernels)
-    assert all(processor.has_learned_registers for processor in installed)
-    assert not any(processor.preserve_learned_register_mask for processor in installed)


-def test_ltx25_connector_attention_preserves_learned_register_mask(monkeypatch):
-    installed = []
-
-    class OmniAttention:
-        def __init__(self, **_kwargs):
-            pass
-
-    class Attention:
-        heads = 32
-        head_dim = 128
-        inner_kv_dim = 4096
-
-        def set_processor(self, processor):
-            installed.append(processor)
-
-    monkeypatch.setattr(ltx2_components, "OmniAttention", OmniAttention)
-    connectors = SimpleNamespace(
-        video_connector=SimpleNamespace(
-            learnable_registers=object(),
-            transformer_blocks=[SimpleNamespace(attn1=Attention())],
-        ),
-        audio_connector=SimpleNamespace(
-            learnable_registers=object(),
-            transformer_blocks=[SimpleNamespace(attn1=Attention())],
-        ),
-    )
-
-    _install_connector_attention(connectors, preserve_learned_register_mask=True)
-
-    assert all(processor.has_learned_registers for processor in installed)
-    assert all(processor.preserve_learned_register_mask for processor in installed)
-
-
-def test_ltx_connector_attention_dispatches_through_omni_kernel():
+@pytest.mark.parametrize("has_learned_registers", [False, True])
+def test_ltx_connector_attention_dispatches_through_omni_kernel(has_learned_registers):
    calls = []

    class Backend:
@@ -775,16 +732,12 @@ def test_ltx_connector_attention_dispatches_through_omni_kernel():
            calls.append((query, key, value, metadata))
            return query

-    class UpcastingNorm(torch.nn.Module):
-        def forward(self, tensor):
-            return tensor.float()
-
    attention = SimpleNamespace(
        heads=2,
        head_dim=4,
        inner_kv_dim=8,
-        norm_q=UpcastingNorm(),
-        norm_k=UpcastingNorm(),
+        norm_q=torch.nn.Identity(),
+        norm_k=torch.nn.Identity(),
        to_q=torch.nn.Identity(),
        to_k=torch.nn.Identity(),
        to_v=torch.nn.Identity(),
@@ -793,29 +746,24 @@ def test_ltx_connector_attention_dispatches_through_omni_kernel():
        rope_type="split",
        omni_attention=OmniAttention(),
    )
-    hidden_states = torch.randn(2, 3, 8, dtype=torch.bfloat16)
+    hidden_states = torch.randn(2, 3, 8)
    additive_mask = torch.zeros(2, 1, 3, 3)
-    rotary = (
-        torch.randn(2, 2, 3, 2, dtype=torch.float32),
-        torch.randn(2, 2, 3, 2, dtype=torch.float32),
-    )

-    output = ltx2_components._LTXConnectorAttnProcessor()(
+    output = ltx2_components._LTXConnectorAttnProcessor(has_learned_registers=has_learned_registers)(
        attention,
        hidden_states,
        attention_mask=additive_mask,
-        query_rotary_emb=rotary,
    )

    query, key, value, metadata = calls[0]
    assert query.shape == key.shape == value.shape == (2, 3, 2, 4)
-    assert query.dtype == key.dtype == value.dtype == torch.bfloat16
-    assert metadata.attn_mask.shape == (2, 3)
-    assert metadata.attn_mask.dtype == torch.bool
-    assert metadata.attn_mask.all()
-    expected_rotary = tuple(component.to(torch.bfloat16) for component in rotary)
-    expected = ltx2_components.apply_split_rotary_emb(hidden_states, expected_rotary, head_dim=4)
-    torch.testing.assert_close(output, expected, rtol=0, atol=0)
+    if has_learned_registers:
+        assert metadata is None
+    else:
+        assert metadata.attn_mask.shape == (2, 3)
+        assert metadata.attn_mask.dtype == torch.bool
+        assert metadata.attn_mask.all()
+    torch.testing.assert_close(output, hidden_states)


 def test_ltx2_distilled_pipeline_shares_recipe_runtime():
@@ -1759,7 +1707,6 @@ class TestLTXForwardStages:
        assert seen["request_inputs"].output_type == "latent"
        assert seen["kwargs"] == {
            "request_sigmas": request_sigmas,
-            "request_phase_sigmas": None,
            "image": None,
        }

@@ -1877,16 +1824,6 @@ class TestRegistryIntegration:
            "pipeline_ltx2_two_stage",
            "LTX2TwoStagePipeline",
        )
-        assert _DIFFUSION_MODELS["LTX2DistilledOneStagePipeline"] == (
-            "ltx2",
-            "pipeline_ltx2",
-            "LTX2DistilledOneStagePipeline",
-        )
-        assert _DIFFUSION_MODELS["LTX2DistilledTwoStagePipeline"] == (
-            "ltx2",
-            "pipeline_ltx2_two_stage",
-            "LTX2DistilledTwoStagePipeline",
-        )
        assert _DIFFUSION_MODELS["LTX2DistilledPipeline"] == (
            "ltx2",
            "pipeline_ltx2_two_stage",
@@ -1911,8 +1848,6 @@ class TestRegistryIntegration:
        expected = [
            "LTX2Pipeline",
            "LTX2TwoStagePipeline",
-            "LTX2DistilledOneStagePipeline",
-            "LTX2DistilledTwoStagePipeline",
            "LTX2DistilledPipeline",
            "LTX2T2VDMD2Pipeline",
            "LTX2I2VDMD2Pipeline",
@@ -1926,8 +1861,6 @@ class TestRegistryIntegration:
        [
            "LTX2Pipeline",
            "LTX2TwoStagePipeline",
-            "LTX2DistilledOneStagePipeline",
-            "LTX2DistilledTwoStagePipeline",
            "LTX2DistilledPipeline",
            "LTX2T2VDMD2Pipeline",
            "LTX2I2VDMD2Pipeline",
@@ -7,54 +7,18 @@ import pytest
 import torch
 from torch import nn

+from vllm_omni.diffusion.attention.backends.abstract import AttentionMetadata
 from vllm_omni.diffusion.cache.cachedit import CacheDiTAdapterConfig
-from vllm_omni.diffusion.models.ltx2 import ltx2_transformer
 from vllm_omni.diffusion.models.ltx2.ltx2_transformer import (
-    LTX2FeedForward,
+    LTX2AudioVideoAttnProcessor,
    LTX2VideoTransformer3DModel,
-    LTX2VideoTransformerBlock,
+    _LTX2ParallelAttention,
    _make_rms_norm,
-    apply_interleaved_rotary_emb,
-    apply_keyframes_absolute_embedding,
-    apply_split_rotary_emb,
 )

 pytestmark = [pytest.mark.core_model, pytest.mark.diffusion, pytest.mark.cpu]


-def test_ltx_interleaved_rope_matches_official_model_dtype_arithmetic():
-    torch.manual_seed(0)
-    x = torch.randn(2, 3, 8, dtype=torch.bfloat16)
-    cos = torch.randn(2, 3, 8, dtype=x.dtype)
-    sin = torch.randn(2, 3, 8, dtype=x.dtype)
-    x_real, x_imag = x.unflatten(2, (4, 2)).unbind(-1)
-    rotated = torch.stack([-x_imag, x_real], dim=-1).flatten(2)
-    expected = x * cos + rotated * sin
-
-    actual = apply_interleaved_rotary_emb(x, (cos, sin))
-
-    torch.testing.assert_close(actual, expected, rtol=0, atol=0)
-
-
-def test_ltx_split_rope_matches_official_model_dtype_arithmetic():
-    torch.manual_seed(1)
-    x = torch.randn(2, 3, 8, dtype=torch.bfloat16)
-    cos = torch.randn(2, 3, 4, dtype=x.dtype)
-    sin = torch.randn(2, 3, 4, dtype=x.dtype)
-    split_x = x.reshape(2, 3, 2, 4)
-    first_x, second_x = split_x.chunk(2, dim=-2)
-    cos_u = cos.unsqueeze(-2)
-    sin_u = sin.unsqueeze(-2)
-    expected = split_x * cos_u
-    expected[..., :1, :].addcmul_(-sin_u, second_x)
-    expected[..., 1:, :].addcmul_(sin_u, first_x)
-    expected = expected.reshape_as(x)
-
-    actual = apply_split_rotary_emb(x, (cos, sin), head_dim=8)
-
-    torch.testing.assert_close(actual, expected, rtol=0, atol=0)
-
-
 def test_ltx_rms_norm_no_affine_matches_official_torch_module():
    norm = _make_rms_norm(8, eps=1e-6, elementwise_affine=False)

@@ -72,53 +36,11 @@ def test_ltx_rms_norm_affine_weight_remains_parameter():
    assert "weight" in norm.state_dict()


-def test_ltx_attention_restores_projection_dtype_after_upcasting_qk_norms():
-    calls = []
-
-    class UpcastingNorm(nn.Module):
-        def forward(self, tensor):
-            return tensor.float()
-
-    class RecordingAttention:
-        def __call__(self, query, key, value, metadata):
-            calls.append((query, key, value, metadata))
-            return query
-
-    attention = SimpleNamespace(
-        heads=2,
-        head_dim=4,
-        to_qkv=None,
-        to_q=nn.Identity(),
-        to_k=nn.Identity(),
-        to_v=nn.Identity(),
-        norm_q=UpcastingNorm(),
-        norm_k=UpcastingNorm(),
-        to_gate_logits=None,
-        attn=RecordingAttention(),
-        to_out=(nn.Identity(), nn.Identity()),
-        rope_type="split",
-    )
-    hidden_states = torch.randn(2, 3, 8, dtype=torch.bfloat16)
-    rotary = (
-        torch.ones(2, 2, 3, 2, dtype=torch.float32),
-        torch.zeros(2, 2, 3, 2, dtype=torch.float32),
-    )
-
-    processor = ltx2_transformer.LTX2AudioVideoAttnProcessor()
-    processor._slice_rope_for_tp = lambda rope, _: rope
-    output = processor(attention, hidden_states, query_rotary_emb=rotary)
-
-    query, key, value, metadata = calls[0]
-    assert query.dtype == key.dtype == value.dtype == torch.bfloat16
-    assert metadata is None
-    torch.testing.assert_close(output, hidden_states)
-
-
-def test_ltx_transformer_has_fused_cfg_cache_dit_config():
+def test_ltx_transformer_has_separate_cfg_cache_dit_config():
    adapter_config = getattr(LTX2VideoTransformer3DModel, "_cache_dit_adapter_config")

    assert isinstance(adapter_config, CacheDiTAdapterConfig)
-    assert not adapter_config.has_separate_cfg
+    assert adapter_config.has_separate_cfg


 def test_ltx_transformer_exposes_hsdp_shard_conditions_for_blocks():
@@ -154,9 +76,8 @@ def test_ltx_sp_plan_shards_video_and_audio_timesteps_together(rope_type):


 def test_ltx_sp_keeps_cross_attention_key_padding_mask(monkeypatch):
-    processor_cls = ltx2_transformer.LTX2AudioVideoAttnProcessor
-    monkeypatch.setattr(processor_cls, "_is_sp_enabled", staticmethod(lambda: True))
-    processor = processor_cls()
+    monkeypatch.setattr(LTX2AudioVideoAttnProcessor, "_is_sp_enabled", staticmethod(lambda: True))
+    processor = LTX2AudioVideoAttnProcessor()
    hidden_states = torch.zeros(1, 3, 4)
    encoder_hidden_states = torch.zeros(1, 4, 4)
    key_padding_mask = torch.tensor([[True, True, True, False]])
@@ -192,7 +113,7 @@ def _fake_ltx_parallel_attention():

        return forward

-    layer = object.__new__(ltx2_transformer._LTX2ParallelAttention)
+    layer = object.__new__(_LTX2ParallelAttention)
    nn.Module.__init__(layer)
    layer.attn_backend = Backend
    layer.attention = SimpleNamespace(forward=run("native"))
@@ -202,281 +123,33 @@ def _fake_ltx_parallel_attention():


@pytest.mark.parametrize(("query_length", "key_length"), [(4, 4), (3, 4)])
-def test_ltx_masked_attention_uses_sdpa(query_length, key_length):
+def test_ltx_masked_cross_attention_uses_sdpa(query_length, key_length):
    layer, calls = _fake_ltx_parallel_attention()
    query = torch.zeros(1, query_length, 2, 4, dtype=torch.bfloat16)
    key = value = torch.zeros(1, key_length, 2, 4, dtype=torch.bfloat16)
    key_padding_mask = torch.tensor([[True] * (key_length - 1) + [False]])
-    metadata = ltx2_transformer.AttentionMetadata(attn_mask=key_padding_mask[:, None, None, :])
+    metadata = AttentionMetadata(attn_mask=key_padding_mask[:, None, None, :])

    layer._run_local_attention(query, key, value, metadata)

    assert calls == ["sdpa"]


-def test_ltx_unmasked_attention_keeps_native_backend():
+def test_ltx_masked_self_attention_uses_sdpa():
+    layer, calls = _fake_ltx_parallel_attention()
+    query = key = value = torch.zeros(1, 4, 2, 4, dtype=torch.bfloat16)
+    key_padding_mask = torch.tensor([[True, True, True, False]])
+    metadata = AttentionMetadata(attn_mask=key_padding_mask[:, None, None, :])
+
+    layer._run_local_attention(query, key, value, metadata)
+
+    assert calls == ["sdpa"]
+
+
+def test_ltx_unmasked_cross_attention_keeps_native_backend():
    layer, calls = _fake_ltx_parallel_attention()
    query = key = value = torch.zeros(1, 4, 2, 4, dtype=torch.bfloat16)

    layer._run_local_attention(query, key, value, None)

    assert calls == ["native"]
-
-
-def test_ltx25_feed_forward_bias_flag_controls_both_linears(monkeypatch):
-    class FakeParallelLinear(nn.Linear):
-        def __init__(self, dim_in, dim_out, *, bias=True, **kwargs):
-            super().__init__(dim_in, dim_out, bias=bias)
-
-    monkeypatch.setattr(ltx2_transformer, "ColumnParallelLinear", FakeParallelLinear)
-    monkeypatch.setattr(ltx2_transformer, "RowParallelLinear", FakeParallelLinear)
-
-    without_bias = LTX2FeedForward(8, inner_dim=16, bias=False)
-    with_bias = LTX2FeedForward(8, inner_dim=16, bias=True)
-
-    assert without_bias.net[0].proj.bias is None
-    assert without_bias.net[2].bias is None
-    assert with_bias.net[0].proj.bias is not None
-    assert with_bias.net[2].bias is not None
-
-
-def _build_tiny_ltx_transformer(monkeypatch, **overrides):
-    class FakeBlock(nn.Module):
-        def __init__(self, **kwargs):
-            super().__init__()
-            self.init_kwargs = kwargs
-
-        def forward(self, hidden_states, audio_hidden_states, *_args, **_kwargs):
-            return hidden_states, audio_hidden_states
-
-    monkeypatch.setattr(ltx2_transformer, "LTX2VideoTransformerBlock", FakeBlock)
-    kwargs = {
-        "in_channels": 4,
-        "out_channels": 4,
-        "num_attention_heads": 2,
-        "attention_head_dim": 4,
-        "cross_attention_dim": 8,
-        "vae_scale_factors": (1, 1, 1),
-        "pos_embed_max_pos": 4,
-        "base_height": 8,
-        "base_width": 8,
-        "audio_in_channels": 4,
-        "audio_out_channels": 4,
-        "audio_num_attention_heads": 2,
-        "audio_attention_head_dim": 4,
-        "audio_cross_attention_dim": 8,
-        "audio_scale_factor": 1,
-        "audio_pos_embed_max_pos": 4,
-        "audio_sampling_rate": 16,
-        "audio_hop_length": 4,
-        "num_layers": 1,
-        "caption_channels": 8,
-        "use_prompt_embeddings": False,
-        "cross_attn_mod": True,
-        "audio_cross_attn_mod": True,
-    }
-    kwargs.update(overrides)
-    return LTX2VideoTransformer3DModel(**kwargs)
-
-
-def test_ltx25_keyframe_absolute_embedding_matches_official():
-    hidden_states = torch.arange(32, dtype=torch.bfloat16).view(1, 4, 8)
-    keyframes_mask = torch.tensor([[[1.0], [1.0], [0.0], [0.0]]])
-    embedding = torch.linspace(-0.5, 0.5, 8, dtype=torch.bfloat16).view(1, 8)
-
-    actual = apply_keyframes_absolute_embedding(hidden_states, keyframes_mask, embedding)
-    expected = hidden_states + (keyframes_mask > 0).to(torch.bfloat16) * embedding
-
-    torch.testing.assert_close(actual, expected, rtol=0, atol=0)
-    assert apply_keyframes_absolute_embedding(hidden_states, None, embedding) is hidden_states
-
-
-def test_ltx25_constructor_flags_and_native_weight_loader(monkeypatch):
-    legacy_model = _build_tiny_ltx_transformer(monkeypatch)
-
-    assert legacy_model.config.ff_bias is True
-    assert legacy_model.config.audio_ff_bias is True
-    assert legacy_model.config.use_prompt_adaln_single is True
-    assert legacy_model.config.use_keyframes_abs_pos_embedding is False
-    assert hasattr(legacy_model, "prompt_adaln")
-    assert hasattr(legacy_model, "audio_prompt_adaln")
-    assert not hasattr(legacy_model, "keyframes_abs_pos_embedding")
-    assert legacy_model.transformer_blocks[0].init_kwargs["ff_bias"] is True
-    assert legacy_model.transformer_blocks[0].init_kwargs["audio_ff_bias"] is True
-
-    ltx25_model = _build_tiny_ltx_transformer(
-        monkeypatch,
-        ff_bias=False,
-        audio_ff_bias=True,
-        use_prompt_adaln_single=False,
-        use_keyframes_abs_pos_embedding=True,
-    )
-
-    assert ltx25_model.keyframes_abs_pos_embedding.shape == (1, 8)
-    torch.testing.assert_close(ltx25_model.keyframes_abs_pos_embedding, torch.zeros(1, 8))
-    assert not hasattr(ltx25_model, "prompt_adaln")
-    assert not hasattr(ltx25_model, "audio_prompt_adaln")
-    assert ltx25_model.transformer_blocks[0].init_kwargs["ff_bias"] is False
-    assert ltx25_model.transformer_blocks[0].init_kwargs["audio_ff_bias"] is True
-
-    monkeypatch.setattr(ltx2_transformer, "get_tensor_model_parallel_world_size", lambda: 1)
-    expected = torch.arange(8, dtype=torch.float32).view(1, 8)
-    loaded = ltx25_model.load_weights([("keyframes_abs_pos_embedding", expected)])
-
-    assert loaded == {"keyframes_abs_pos_embedding"}
-    torch.testing.assert_close(ltx25_model.keyframes_abs_pos_embedding, expected)
-
-
-def test_ltx_cross_attention_modulations_use_official_timesteps(monkeypatch):
-    class CaptureAda(nn.Module):
-        def __init__(self):
-            super().__init__()
-            self.calls = []
-
-        def forward(self, timestep, *, batch_size, hidden_dtype):
-            self.calls.append(timestep.detach().clone())
-            embedding = torch.zeros(timestep.numel(), 8, dtype=hidden_dtype)
-            return embedding, embedding
-
-    class CaptureRope(nn.Module):
-        def __init__(self):
-            super().__init__()
-            self.output_dtypes = []
-
-        def forward(self, _coords, *, device, out_dtype=None):
-            self.output_dtypes.append(out_dtype)
-            empty = torch.empty(0, device=device, dtype=out_dtype)
-            return empty, empty
-
-    model = _build_tiny_ltx_transformer(monkeypatch)
-    rope = CaptureRope()
-    model.rope = model.audio_rope = model.cross_attn_rope = model.cross_attn_audio_rope = rope
-    model.prompt_modulation = False
-    video_scale_shift = CaptureAda()
-    video_gate = CaptureAda()
-    audio_scale_shift = CaptureAda()
-    audio_gate = CaptureAda()
-    model.av_cross_attn_video_scale_shift = video_scale_shift
-    model.av_cross_attn_video_a2v_gate = video_gate
-    model.av_cross_attn_audio_scale_shift = audio_scale_shift
-    model.av_cross_attn_audio_v2a_gate = audio_gate
-
-    model(
-        hidden_states=torch.zeros(1, 2, 4),
-        audio_hidden_states=torch.zeros(1, 1, 4),
-        encoder_hidden_states=torch.zeros(1, 1, 8),
-        audio_encoder_hidden_states=torch.zeros(1, 1, 8),
-        timestep=torch.tensor([[0.0, 2.0]]),
-        audio_timestep=torch.tensor([[5.0]]),
-        sigma=torch.tensor([2.0]),
-        audio_sigma=torch.tensor([5.0]),
-        video_coords=torch.zeros(1, 3, 2, 2),
-        audio_coords=torch.zeros(1, 1, 1, 2),
-    )
-
-    torch.testing.assert_close(video_scale_shift.calls[0], torch.tensor([0.0, 2.0]))
-    torch.testing.assert_close(video_gate.calls[0], torch.tensor([5.0]))
-    torch.testing.assert_close(audio_scale_shift.calls[0], torch.tensor([5.0]))
-    torch.testing.assert_close(audio_gate.calls[0], torch.tensor([2.0]))
-    assert len(rope.output_dtypes) == 4
-    assert all(dtype is torch.float32 for dtype in rope.output_dtypes)
-
-
-def test_ltx25_static_prompt_table_is_used_without_prompt_adaln():
-    class CaptureAttention(nn.Module):
-        def __init__(self):
-            super().__init__()
-            self.encoder_hidden_states = None
-
-        def forward(self, hidden_states, encoder_hidden_states=None, **kwargs):
-            if encoder_hidden_states is not None:
-                self.encoder_hidden_states = encoder_hidden_states.detach().clone()
-            return torch.zeros_like(hidden_states)
-
-    class ZeroOutput(nn.Module):
-        def forward(self, hidden_states):
-            return torch.zeros_like(hidden_states)
-
-    block = object.__new__(LTX2VideoTransformerBlock)
-    nn.Module.__init__(block)
-    block.video_cross_attn_adaln = True
-    block.audio_cross_attn_adaln = True
-    block.cross_attn_adaln = True
-    block.scale_shift_table = nn.Parameter(torch.zeros(9, 2))
-    block.audio_scale_shift_table = nn.Parameter(torch.zeros(9, 2))
-    block.prompt_scale_shift_table = nn.Parameter(torch.tensor([[1.0, 2.0], [3.0, 4.0]]))
-    block.audio_prompt_scale_shift_table = nn.Parameter(torch.tensor([[0.5, 1.0], [1.5, 2.0]]))
-    block.norm1 = nn.Identity()
-    block.audio_norm1 = nn.Identity()
-    block.norm2 = nn.Identity()
-    block.audio_norm2 = nn.Identity()
-    block.norm3 = nn.Identity()
-    block.audio_norm3 = nn.Identity()
-    block.attn1 = CaptureAttention()
-    block.audio_attn1 = CaptureAttention()
-    block.attn2 = CaptureAttention()
-    block.audio_attn2 = CaptureAttention()
-    block.ff = ZeroOutput()
-    block.audio_ff = ZeroOutput()
-
-    hidden_states = torch.ones(1, 1, 2)
-    audio_hidden_states = torch.ones(1, 1, 2)
-    encoder_hidden_states = torch.ones(1, 1, 2)
-    audio_encoder_hidden_states = torch.ones(1, 1, 2)
-    block(
-        hidden_states,
-        audio_hidden_states,
-        encoder_hidden_states,
-        audio_encoder_hidden_states,
-        temb=torch.zeros(1, 1, 18),
-        temb_audio=torch.zeros(1, 1, 18),
-        temb_ca_scale_shift=torch.empty(0),
-        temb_ca_audio_scale_shift=torch.empty(0),
-        temb_ca_gate=torch.empty(0),
-        temb_ca_audio_gate=torch.empty(0),
-        temb_prompt=None,
-        temb_prompt_audio=None,
-        use_a2v_cross_attention=False,
-        use_v2a_cross_attention=False,
-    )
-
-    torch.testing.assert_close(block.attn2.encoder_hidden_states, torch.tensor([[[5.0, 7.0]]]))
-    torch.testing.assert_close(block.audio_attn2.encoder_hidden_states, torch.tensor([[[3.0, 4.0]]]))
-
-
-def test_ltx_gated_attention_uses_tp_sharded_projection(monkeypatch):
-    column_calls = []
-
-    class FakeColumnParallelLinear(nn.Linear):
-        def __init__(self, input_size, output_size, *, bias=True, **kwargs):
-            column_calls.append((output_size, kwargs))
-            super().__init__(input_size, output_size, bias=bias)
-
-    class FakeRowParallelLinear(nn.Linear):
-        def __init__(self, input_size, output_size, *, bias=True, **_kwargs):
-            super().__init__(input_size, output_size, bias=bias)
-
-    class FakeParallelAttention(nn.Module):
-        def __init__(self, **_kwargs):
-            super().__init__()
-
-    monkeypatch.setattr(ltx2_transformer, "ColumnParallelLinear", FakeColumnParallelLinear)
-    monkeypatch.setattr(ltx2_transformer, "RowParallelLinear", FakeRowParallelLinear)
-    monkeypatch.setattr(ltx2_transformer, "_LTX2ParallelAttention", FakeParallelAttention)
-    monkeypatch.setattr(ltx2_transformer, "get_tensor_model_parallel_world_size", lambda: 2)
-
-    attention = ltx2_transformer.LTX2Attention(
-        query_dim=8,
-        heads=4,
-        kv_heads=4,
-        dim_head=2,
-        qk_norm="rms_norm",
-        apply_gated_attention=True,
-        pack_qkv=False,
-    )
-
-    gate_output_size, gate_kwargs = column_calls[-1]
-    assert gate_output_size == 4
-    assert gate_kwargs["gather_output"] is False
-    assert gate_kwargs["return_bias"] is False
-    assert attention.query_num_heads == 2
@@ -5,7 +5,7 @@ import json
 import sys
 from multiprocessing.reduction import ForkingPickler
 from types import SimpleNamespace
-from unittest.mock import Mock, patch
+from unittest.mock import Mock

 import numpy as np
 import pytest
@@ -363,18 +363,8 @@ def test_pipeline_loads_task_selected_dits_and_shared_components_once(
    if expected_partition == "ref2va":
        assert pipeline._transformer_for_task("ref2va") is pipeline.transformer
    assert [source.model_or_path for source in pipeline.weights_sources] == [
-        *(str(tmp_path / partition_name) for partition_name in source_partitions),
-        str(component_path),
+        str(tmp_path / partition_name) for partition_name in source_partitions
    ]
-    assert [source.subfolder for source in pipeline.weights_sources] == [
-        *("transformer" for _ in source_partitions),
-        "text_encoder",
-    ]
-    expected_prefixes = ["transformer."]
-    if expected_dits == 2:
-        expected_prefixes.append("transformers_ref.")
-    expected_prefixes.append("text_encoder.")
-    assert [source.prefix for source in pipeline.weights_sources] == expected_prefixes
    expected_patterns = (
        pipeline_module.MINIMAX_H3_DOWNLOAD_PATTERNS
        if expected_partition == "combined"
@@ -946,66 +936,6 @@ def test_text_encoder_stub_constructs_without_group_or_weights():
    assert list(encoder.named_parameters()) == []


-def test_global_quant_config_is_shared_by_dit_and_encoder():
-    from vllm_omni.quantization import build_quant_config, resolve_component_quant_config
-
-    global_config = build_quant_config("fp8")
-
-    assert resolve_component_quant_config(global_config, "transformer") is global_config
-    assert resolve_component_quant_config(global_config, "text_encoder") is global_config
-
-
-def test_minimax_h3_fp8_component_selection():
-    from vllm_omni.quantization import build_quant_config, resolve_component_quant_config
-
-    dit_only = build_quant_config({"transformer": {"method": "fp8"}})
-    encoder_only = build_quant_config({"text_encoder": {"method": "fp8"}})
-
-    assert resolve_component_quant_config(dit_only, "transformer").get_name() == "fp8"
-    assert resolve_component_quant_config(dit_only, "text_encoder") is None
-    assert resolve_component_quant_config(encoder_only, "transformer") is None
-    assert resolve_component_quant_config(encoder_only, "text_encoder").get_name() == "fp8"
-
-
-def test_text_encoder_rejects_serialized_fp8():
-    from vllm_omni.diffusion.models.minimax_h3.encoder import _validate_encoder_quant_config
-
-    config = SimpleNamespace(
-        get_name=lambda: "fp8",
-        is_checkpoint_fp8_serialized=True,
-    )
-    with pytest.raises(ValueError, match="online FP8 only"):
-        _validate_encoder_quant_config(config)
-
-
-def test_text_encoder_linear_delegates_quantization_to_vllm_factory():
-    from vllm.model_executor.layers.linear import UnquantizedLinearMethod
-
-    import vllm_omni.diffusion.models.minimax_h3.encoder as encoder_module
-
-    group = SimpleNamespace(rank_in_group=0, world_size=1)
-    method = UnquantizedLinearMethod()
-    quant_config = Mock()
-    quant_config.get_quant_method.return_value = method
-    prefix = "text_encoder.text_model.layers.0.mlp.gate_up_proj"
-
-    with (
-        patch("vllm.model_executor.parameter.get_tensor_model_parallel_rank", return_value=0),
-        patch("vllm.model_executor.parameter.get_tensor_model_parallel_world_size", return_value=1),
-    ):
-        linear = encoder_module.MiniMaxH3Qwen3VLMergedColumnParallelLinear(
-            group,
-            input_size=4,
-            intermediate_size=8,
-            dtype=torch.bfloat16,
-            quant_config=quant_config,
-            prefix=prefix,
-        )
-
-    assert linear.quant_method is method
-    quant_config.get_quant_method.assert_called_once_with(linear, prefix=prefix)
-
-
 def test_no_offload_keeps_text_encoder_resident():
    from vllm_omni.diffusion.models.minimax_h3 import MiniMaxH3Pipeline

@@ -1048,21 +978,15 @@ def test_model_offload_uses_hooked_text_encoder_call():
    pipeline.text_encoder.load_to_device.assert_not_called()


-@pytest.mark.parametrize(
-    "offload_flag",
-    ["enable_layerwise_offload", "enable_distributed_layerwise_offload"],
-)
-def test_layerwise_offload_releases_text_encoder(offload_flag):
+def test_layerwise_offload_releases_text_encoder():
    from vllm_omni.diffusion.models.minimax_h3 import MiniMaxH3Pipeline

    pipeline = object.__new__(MiniMaxH3Pipeline)
    torch.nn.Module.__init__(pipeline)
    pipeline.od_config = SimpleNamespace(
        enable_cpu_offload=False,
-        enable_layerwise_offload=False,
-        enable_distributed_layerwise_offload=False,
+        enable_layerwise_offload=True,
    )
-    setattr(pipeline.od_config, offload_flag, True)
    pipeline.text_encoder = Mock()
    expected = torch.ones(2, 3)
    pipeline.text_encoder.encode_ids.return_value = expected
@@ -2064,25 +1988,21 @@ def _one_layer_text_encoder():
    group = _SingleRankEncoderGroup(0)
    layer = nn.Module()
    layer.self_attn = nn.Module()
+    layer.self_attn.qkv_proj = MiniMaxH3Qwen3VLQKVParallelLinear(
+        group,
+        hidden_size=_ENCODER_HIDDEN,
+        num_heads=_ENCODER_NUM_HEADS,
+        num_kv_heads=_ENCODER_NUM_KV_HEADS,
+        head_dim=_ENCODER_HEAD_DIM,
+        dtype=torch.float32,
+    )
    layer.mlp = nn.Module()
-    with (
-        patch("vllm.model_executor.parameter.get_tensor_model_parallel_rank", return_value=0),
-        patch("vllm.model_executor.parameter.get_tensor_model_parallel_world_size", return_value=1),
-    ):
-        layer.self_attn.qkv_proj = MiniMaxH3Qwen3VLQKVParallelLinear(
-            group,
-            hidden_size=_ENCODER_HIDDEN,
-            num_heads=_ENCODER_NUM_HEADS,
-            num_kv_heads=_ENCODER_NUM_KV_HEADS,
-            head_dim=_ENCODER_HEAD_DIM,
-            dtype=torch.float32,
-        )
-        layer.mlp.gate_up_proj = MiniMaxH3Qwen3VLMergedColumnParallelLinear(
-            group,
-            input_size=_ENCODER_HIDDEN,
-            intermediate_size=_ENCODER_INTERMEDIATE,
-            dtype=torch.float32,
-        )
+    layer.mlp.gate_up_proj = MiniMaxH3Qwen3VLMergedColumnParallelLinear(
+        group,
+        input_size=_ENCODER_HIDDEN,
+        intermediate_size=_ENCODER_INTERMEDIATE,
+        dtype=torch.float32,
+    )
    layer.input_layernorm = MiniMaxH3Qwen3VLRMSNorm(_ENCODER_HIDDEN)

    encoder = object.__new__(MiniMaxH3Qwen3VLEncoder)
@@ -2115,7 +2035,6 @@ def _write_text_encoder_checkpoint(path, *, omit=()):
    safetensors.torch.save_file(tensors, str(path / "model.safetensors"))
    index = {"weight_map": dict.fromkeys(tensors, "model.safetensors")}
    (path / "model.safetensors.index.json").write_text(json.dumps(index))
-    return tensors


@pytest.mark.parametrize(
@@ -2137,10 +2056,10 @@ def test_encoder_load_reports_missing_fused_source_shards(
    tmp_path, omitted_sources, expected_detail, unreported_target
 ):
    encoder = _one_layer_text_encoder()
-    weights = _write_text_encoder_checkpoint(tmp_path, omit=omitted_sources)
+    _write_text_encoder_checkpoint(tmp_path, omit=omitted_sources)

    with pytest.raises(RuntimeError) as excinfo:
-        encoder.load_weights(weights.items())
+        encoder._load_weights(str(tmp_path))

    message = str(excinfo.value)
    assert expected_detail in message
@@ -2149,17 +2068,17 @@ def test_encoder_load_reports_missing_fused_source_shards(

 def test_encoder_load_rejects_unloaded_plain_param(tmp_path):
    encoder = _one_layer_text_encoder()
-    weights = _write_text_encoder_checkpoint(tmp_path, omit=("input_layernorm.weight",))
+    _write_text_encoder_checkpoint(tmp_path, omit=("input_layernorm.weight",))

    with pytest.raises(RuntimeError, match=r"weights not loaded.*input_layernorm"):
-        encoder.load_weights(weights.items())
+        encoder._load_weights(str(tmp_path))


 def test_encoder_load_places_every_source_in_its_own_rows(tmp_path):
    encoder = _one_layer_text_encoder()
-    weights = _write_text_encoder_checkpoint(tmp_path)
+    _write_text_encoder_checkpoint(tmp_path)

-    encoder.load_weights(weights.items())
+    encoder._load_weights(str(tmp_path))

    params = dict(encoder.named_parameters())
    q_rows = _ENCODER_NUM_HEADS * _ENCODER_HEAD_DIM
@@ -36,13 +36,7 @@ def test_transformer_declares_cache_sp_layerwise_offload_and_hsdp():
    assert MiniMaxH3DiTModel._layerwise_offload_blocks_attrs == ["blocks"]
    assert MiniMaxH3DiTModel._cache_dit_adapter_config.block_forward_patterns["blocks"] == ForwardPattern.Pattern_3
    assert not MiniMaxH3DiTModel._cache_dit_adapter_config.has_separate_cfg
-    assert set(MiniMaxH3DiTModel._sp_plan) == {
-        "sp_prepare",
-        "local_sp_prepare",
-        "sp_gather",
-    }
-    assert set(MiniMaxH3DiTModel._sp_plan["sp_prepare"]) == {0, 1, 2}
-    assert set(MiniMaxH3DiTModel._sp_plan["local_sp_prepare"]) == {2}
+    assert set(MiniMaxH3DiTModel._sp_plan) == {"sp_prepare", "sp_gather"}

    model = object.__new__(MiniMaxH3DiTModel)
    nn.Module.__init__(model)
@@ -95,279 +89,6 @@ def test_h3_fused_rope_matches_reference_and_preserves_unrotated_dims():
    torch.testing.assert_close(actual[..., 96:], x[..., 96:], atol=0, rtol=0)


-def test_h3_rope_table_materializes_local_rows_in_fused_kernel_layout():
-    from vllm_omni.diffusion.models.minimax_h3.minimax_h3_transformer import (
-        _build_rope_table,
-    )
-
-    local_freqs_half = torch.randn(2, 48)
-    local_freqs = torch.cat((local_freqs_half, local_freqs_half), dim=-1)
-
-    actual = _build_rope_table(local_freqs)
-    expected = torch.cat(
-        (torch.cos(local_freqs_half), torch.sin(local_freqs_half)),
-        dim=-1,
-    ).to(torch.bfloat16)
-
-    assert actual.shape == (2, 96)
-    torch.testing.assert_close(actual, expected, atol=0, rtol=0)
-
-
-def test_strict_sp_local_span_uses_rank_owned_rows(monkeypatch):
-    from vllm_omni.diffusion import forward_context
-    from vllm_omni.diffusion.distributed import parallel_state
-    from vllm_omni.diffusion.models.minimax_h3 import minimax_h3_transformer as h3
-
-    monkeypatch.setattr(forward_context, "is_forward_context_available", lambda: True)
-    monkeypatch.setattr(forward_context, "get_ulysses_mode", lambda **kwargs: "strict")
-    monkeypatch.setattr(parallel_state, "get_sequence_parallel_world_size", lambda: 8)
-    monkeypatch.setattr(parallel_state, "get_sequence_parallel_rank", lambda: 3)
-    monkeypatch.setattr(parallel_state, "get_ulysses_parallel_world_size", lambda: 8)
-    monkeypatch.setattr(parallel_state, "get_ring_parallel_world_size", lambda: 1)
-    monkeypatch.setattr(parallel_state, "get_allgather_parallel_world_size", lambda: 1)
-
-    assert h3._sequence_parallel_local_span(16, hooks_applied=True) == (6, 2)
-
-
-@pytest.mark.parametrize(
-    ("hooks_applied", "mode", "seq_len", "ulysses", "ring", "allgather"),
-    [
-        (False, "strict", 16, 8, 1, 1),
-        (True, "advanced_uaa", 16, 8, 1, 1),
-        (True, "strict", 17, 8, 1, 1),
-        (True, "strict", 16, 4, 2, 1),
-        (True, "strict", 16, 1, 1, 8),
-    ],
-)
-def test_local_span_falls_back_when_local_embedding_is_unsafe(
-    monkeypatch,
-    hooks_applied,
-    mode,
-    seq_len,
-    ulysses,
-    ring,
-    allgather,
-):
-    from vllm_omni.diffusion import forward_context
-    from vllm_omni.diffusion.distributed import parallel_state
-    from vllm_omni.diffusion.models.minimax_h3 import minimax_h3_transformer as h3
-
-    monkeypatch.setattr(forward_context, "is_forward_context_available", lambda: True)
-    monkeypatch.setattr(forward_context, "get_ulysses_mode", lambda **kwargs: mode)
-    monkeypatch.setattr(parallel_state, "get_sequence_parallel_world_size", lambda: 8)
-    monkeypatch.setattr(parallel_state, "get_sequence_parallel_rank", lambda: 3)
-    monkeypatch.setattr(
-        parallel_state,
-        "get_ulysses_parallel_world_size",
-        lambda: ulysses,
-    )
-    monkeypatch.setattr(parallel_state, "get_ring_parallel_world_size", lambda: ring)
-    monkeypatch.setattr(
-        parallel_state,
-        "get_allgather_parallel_world_size",
-        lambda: allgather,
-    )
-
-    assert h3._sequence_parallel_local_span(seq_len, hooks_applied=hooks_applied) == (
-        0,
-        seq_len,
-    )
-
-
-def test_local_embedding_spans_reassemble_multimodal_rows():
-    from vllm_omni.diffusion.models.minimax_h3 import minimax_h3_transformer as h3
-
-    class _IdentityProjection(nn.Module):
-        def forward(self, x):
-            return x, None
-
-    class _IdentityRefiner(nn.Module):
-        def forward(self, x, **kwargs):
-            return x
-
-    model = object.__new__(h3.MiniMaxH3DiTModel)
-    nn.Module.__init__(model)
-    model.hidden_size = 2
-    model.video_patch_proj = _IdentityProjection()
-    model.audio_patch_proj = _IdentityProjection()
-    model.condition_proj = _IdentityProjection()
-    model.token_refiner = _IdentityRefiner()
-    model.time_embedder = nn.Identity()
-
-    seq_len = 16
-    x = torch.arange(seq_len * 2, dtype=torch.float32).reshape(1, seq_len, 2)
-    audio_x = x + 100
-    text = torch.tensor(
-        [[200.0, 201.0], [202.0, 203.0], [204.0, 205.0], [206.0, 207.0]],
-        dtype=torch.bfloat16,
-    )
-    common = {
-        "x": x,
-        "audio_x": audio_x,
-        "text_embeddings_selected": text,
-        "unique_timesteps": torch.tensor([0.2, 0.3]),
-        "img_pos": torch.tensor([1, 2, 6, 7, 11, 12]),
-        "audio_pos": torch.tensor([3, 4, 8, 9, 13, 14]),
-        "text_pos": torch.tensor([0, 5, 10, 15]),
-        "refiner_cu_seqlens": torch.tensor([0, 4, 4], dtype=torch.int32),
-        "refiner_max_seqlen": 4,
-        "seq_len": seq_len,
-        "device": torch.device("cpu"),
-        "local_span": (0, seq_len),
-    }
-
-    full, full_t_emb = model._embed(**common)
-    local_spans = []
-    for rank in range(8):
-        local_span = (rank * (seq_len // 8), seq_len // 8)
-        local, local_t_emb = model._embed(
-            **{
-                **common,
-                "local_span": local_span,
-            }
-        )
-        assert local.shape == (2, 2)
-        torch.testing.assert_close(local_t_emb, full_t_emb)
-        local_spans.append(local)
-
-    torch.testing.assert_close(torch.cat(local_spans), full, atol=0, rtol=0)
-
-
-@pytest.mark.parametrize(
-    "local_projection",
-    [False, True],
-    ids=["fallback", "strict-local"],
-)
-def test_sp_output_boundary_preserves_logits_and_minimizes_gather(
-    monkeypatch,
-    local_projection,
-):
-    from vllm_omni.diffusion.models.minimax_h3 import minimax_h3_transformer as h3
-
-    class _StaticRope(nn.Module):
-        input_rows = None
-
-        def forward(self, position_ids):
-            self.input_rows = position_ids.shape[1]
-            return torch.zeros(position_ids.shape[1], 6)
-
-    class _Prepare(nn.Module):
-        def __init__(self, shard):
-            super().__init__()
-            self.shard = shard
-
-        def forward(self, hidden, rope, combined):
-            local_len = hidden.shape[0] // 2 if self.shard else hidden.shape[0]
-            return hidden[:local_len], rope[:local_len], combined[:local_len]
-
-    class _Gather(nn.Module):
-        def __init__(self, remote_rows, replace):
-            super().__init__()
-            self.remote_rows = remote_rows
-            self.replace = replace
-            self.input_width = None
-
-        def forward(self, tensor):
-            self.input_width = tensor.shape[-1]
-            if self.replace:
-                return self.remote_rows
-            return torch.cat((tensor, self.remote_rows), dim=0)
-
-    class _FinalLayer(nn.Module):
-        def forward(self, hidden, *, t_emb, inverse_indices):
-            del t_emb
-            inverse = inverse_indices.to(torch.float32).unsqueeze(-1)
-            return (
-                hidden[:, :2].to(torch.float32) + inverse,
-                hidden[:, 2:3].to(torch.float32) - inverse,
-            )
-
-    seq_len = 8
-    full_hidden = torch.arange(seq_len * 4, dtype=torch.bfloat16).reshape(seq_len, 4)
-    inverse_indices = torch.tensor([0, 1, 0, 1, 1, 0, 1, 0])
-    final_layer = _FinalLayer()
-    remote_video, remote_audio = final_layer(
-        full_hidden[seq_len // 2 :],
-        t_emb=torch.empty(0),
-        inverse_indices=inverse_indices[seq_len // 2 :],
-    )
-    if local_projection:
-        remote_rows = torch.cat((remote_video, remote_audio), dim=-1)
-    else:
-        remote_rows = full_hidden
-    gather = _Gather(remote_rows, replace=not local_projection)
-
-    model = object.__new__(h3.MiniMaxH3DiTModel)
-    nn.Module.__init__(model)
-    model.arch = h3.MiniMaxH3DiTArchConfig(
-        latents_dim=2,
-        audio_latents_dim=1,
-        patch_size=(1, 1, 1),
-    )
-    rope = _StaticRope()
-    model.rope = rope
-    model.sp_prepare = _Prepare(shard=True)
-    model.local_sp_prepare = _Prepare(shard=False)
-    model.sp_gather = gather
-    model.blocks = nn.ModuleList()
-    model.final_layer = final_layer
-    monkeypatch.setattr(
-        model,
-        "_embed",
-        lambda **kwargs: (
-            full_hidden[: seq_len // 2] if local_projection else full_hidden,
-            torch.tensor([[0.5]], dtype=torch.float32),
-        ),
-    )
-
-    def _local_span(*args, **kwargs):
-        del args, kwargs
-        return (0, seq_len // 2) if local_projection else (0, seq_len)
-
-    monkeypatch.setattr(
-        h3,
-        "_sequence_parallel_local_span",
-        _local_span,
-    )
-
-    img_pos = torch.tensor([0, 2, 4, 6])
-    audio_pos = torch.tensor([1, 3, 5, 7])
-    video, audio = model(
-        x=torch.zeros(1, seq_len, 2),
-        audio_x=torch.zeros(1, seq_len, 1),
-        img_position_ids=torch.zeros(1, seq_len, 3, dtype=torch.long),
-        unique_timesteps=torch.tensor([0.5]),
-        inverse_indices=inverse_indices,
-        update_mask=torch.ones(img_pos.numel()),
-        token_tags=torch.zeros(seq_len, dtype=torch.long),
-        prompt_embeds=torch.empty(0, 2),
-        img_pos_info={"position_ids": img_pos},
-        audio_pos_info={"position_ids": audio_pos},
-        text_pos_info={"position_ids": torch.empty(0, dtype=torch.long)},
-        img_pos_for_infer_output_info={"position_ids": img_pos},
-        packed_seq_params={
-            "cu_seqlens_q": torch.tensor([0, seq_len, seq_len], dtype=torch.int32),
-            "max_seqlen_q": seq_len,
-        },
-        refiner_packed_seq_params={
-            "cu_seqlens_q": torch.tensor([0, 0, 0], dtype=torch.int32),
-            "max_seqlen_q": 0,
-        },
-    )
-
-    expected_video, expected_audio = final_layer(
-        full_hidden,
-        t_emb=torch.empty(0),
-        inverse_indices=inverse_indices,
-    )
-    torch.testing.assert_close(video, expected_video.index_select(0, img_pos))
-    torch.testing.assert_close(audio, expected_audio.index_select(0, audio_pos))
-    expected_gather_width = 3 if local_projection else full_hidden.shape[-1]
-    expected_rope_rows = seq_len // 2 if local_projection else seq_len
-    assert gather.input_width == expected_gather_width
-    assert rope.input_rows == expected_rope_rows
-
-
@pytest.mark.parametrize(
    ("tp_size", "message"),
    [
@@ -191,49 +191,6 @@ def test_model_load_weights_transforms_before_calling_vllm_loader():
    ]


-def test_loader_adapter_declares_equivalent_direct_mmap_transform(monkeypatch):
-    from vllm_omni.diffusion.model_loader.checkpoint_adapters import (
-        get_direct_mmap_adapter,
-    )
-    from vllm_omni.diffusion.models.minimax_h3 import minimax_h3_transformer as h3
-
-    monkeypatch.setattr(h3, "QKVParallelLinear", _FakeLinear)
-    monkeypatch.setattr(h3, "RowParallelLinear", _FakeLinear)
-    monkeypatch.setattr(h3, "Attention", _FakeAttention)
-
-    arch = h3.MiniMaxH3DiTArchConfig(
-        hidden_size=1,
-        num_attention_heads=2,
-        attention_head_dim=1,
-        rope_inv_freq_len=1,
-    )
-    attention = h3.MiniMaxH3Attention(
-        arch,
-        quant_config=None,
-        prefix="blocks.0.attn",
-    )
-    transformer = object.__new__(h3.MiniMaxH3DiTModel)
-    nn.Module.__init__(transformer)
-    transformer.blocks = nn.ModuleList([nn.Module()])
-    transformer.blocks[0].attn = attention
-    pipeline = nn.Module()
-    pipeline.transformer = transformer
-
-    adapter = get_direct_mmap_adapter(pipeline)
-    assert adapter is not None
-    policy = adapter.policy_for(
-        "transformer.blocks.0.attn.qkv_proj.weight",
-        attention.qkv_proj.weight,
-    )
-    assert policy is not None
-    assert policy.allow_custom_loader
-    assert policy.transform is not None
-    checkpoint_weight = torch.arange(6, dtype=torch.float32).reshape(6, 1)
-
-    assert policy.transform(checkpoint_weight)[:, 0].tolist() == [0, 3, 1, 4, 2, 5]
-    assert not hasattr(attention.qkv_proj.weight, "mmap_weight_transform")
-
-
 def test_pipeline_resolves_transformer_component_quant_config():
    from vllm_omni.diffusion.models.minimax_h3.pipeline_minimax_h3 import (
        _resolve_component_quant_config,
@@ -254,18 +211,3 @@ def test_pipeline_resolves_transformer_component_quant_config():
    assert transformer_config.ignored_layers == ignored_layers
    assert _resolve_component_quant_config(component_config, "transformer") is transformer_config
    assert _resolve_component_quant_config(transformer_config, "transformer") is transformer_config
-
-
-def test_pipeline_strips_prequantized_text_encoder_config():
-    from vllm_omni.diffusion.models.minimax_h3.pipeline_minimax_h3 import _resolve_minimax_h3_text_encoder_quant_config
-    from vllm_omni.quantization import ComponentQuantizationConfig
-
-    prequantized = Mock()
-    prequantized.get_name.return_value = "modelopt"
-    online_fp8 = Mock()
-    online_fp8.get_name.return_value = "fp8"
-    component_config = ComponentQuantizationConfig({"text_encoder": prequantized})
-
-    assert _resolve_minimax_h3_text_encoder_quant_config(online_fp8) is online_fp8
-    assert _resolve_minimax_h3_text_encoder_quant_config(prequantized) is None
-    assert _resolve_minimax_h3_text_encoder_quant_config(component_config) is None
@@ -142,6 +142,12 @@ def _generate_joint_output(omni, config: QualityTestConfig):

    first = outputs[0]
    worker_peak_memory_mb = float(getattr(first, "peak_memory_mb", 0.0) or 0.0)
+    if hasattr(first, "request_output") and isinstance(first.request_output, list):
+        first = first.request_output[0]
+        worker_peak_memory_mb = max(
+            worker_peak_memory_mb,
+            float(getattr(first, "peak_memory_mb", 0.0) or 0.0),
+        )
    if not hasattr(first, "images") or not first.images:
        raise ValueError("Could not extract video frames from H3 output")
    frames = first.images[0]
@@ -8,28 +8,15 @@ import torch
 from PIL import Image
 from torch import nn

-from tests.diffusion.models.wan2_2.conftest import StubScheduler, StubTransformer, StubVAE, noop_progress_bar
+from tests.diffusion.models.wan2_2.conftest import StubTransformer, StubVAE, noop_progress_bar
 from vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_i2v import (
    Wan22I2VPipeline,
-    get_wan22_i2v_post_process_func,
    get_wan22_i2v_pre_process_func,
 )
-from vllm_omni.diffusion.worker.request_batch import DiffusionRequestBatch

 pytestmark = [pytest.mark.core_model, pytest.mark.cpu, pytest.mark.diffusion]


-def test_wan22_i2v_postprocess_honors_request_output_type() -> None:
-    video = torch.zeros(1, 4, 1, 2, 2)
-
-    output = get_wan22_i2v_post_process_func(SimpleNamespace())(
-        video,
-        sampling_params=SimpleNamespace(output_type="latent"),
-    )
-
-    assert output is video
-
-
 def _make_i2v_pipeline(*, expand_timesteps: bool) -> Wan22I2VPipeline:
    pipeline = object.__new__(Wan22I2VPipeline)
    nn.Module.__init__(pipeline)
@@ -44,29 +31,6 @@ def _make_i2v_pipeline(*, expand_timesteps: bool) -> Wan22I2VPipeline:
    return pipeline


-def _make_i2v_sampling(**overrides):
-    values = {
-        "height": 16,
-        "width": 16,
-        "num_frames": 5,
-        "num_inference_steps": 1,
-        "guidance_scale_provided": True,
-        "guidance_scale": 1.0,
-        "guidance_scale_2": None,
-        "guidance_scale_2_provided": False,
-        "boundary_ratio": None,
-        "generator": None,
-        "seed": None,
-        "num_outputs_per_prompt": 1,
-        "max_sequence_length": 8,
-        "latents": None,
-        "output_type": "latent",
-        "extra_args": {},
-    }
-    values.update(overrides)
-    return SimpleNamespace(**values)
-
-
 def test_i2v_preprocess_requires_image_and_resizes_to_480p_aspect() -> None:
    preprocess = get_wan22_i2v_pre_process_func(SimpleNamespace())
    request = SimpleNamespace(
@@ -89,88 +53,6 @@ def test_i2v_preprocess_requires_image_and_resizes_to_480p_aspect() -> None:
        preprocess(missing_image)


-def _make_i2v_preprocess_request(last_image):
-    return SimpleNamespace(
-        prompt={
-            "prompt": "p",
-            "multi_modal_data": {
-                "image": Image.new("RGB", (320, 160), "red"),
-                "last_image": last_image,
-            },
-        },
-        sampling_params=SimpleNamespace(height=16, width=16),
-    )
-
-
-def test_i2v_preprocess_treats_empty_last_image_list_as_absent() -> None:
-    preprocess = get_wan22_i2v_pre_process_func(SimpleNamespace())
-
-    result = preprocess(_make_i2v_preprocess_request([]))
-
-    assert result.prompt["multi_modal_data"]["last_image"] is None
-    assert result.batch_compatibility_key == ("wan22_i2v_last_image", False)
-
-
-def test_i2v_preprocess_unwraps_single_last_image_list() -> None:
-    preprocess = get_wan22_i2v_pre_process_func(SimpleNamespace())
-    last_image = Image.new("RGB", (16, 16), "blue")
-
-    result = preprocess(_make_i2v_preprocess_request([last_image]))
-
-    assert result.prompt["multi_modal_data"]["last_image"] is last_image
-    assert result.batch_compatibility_key == ("wan22_i2v_last_image", True)
-
-
-@pytest.mark.parametrize(
-    "last_image",
-    [
-        Image.new("RGB", (16, 16), "blue"),
-        torch.zeros(3, 16, 16),
-    ],
-)
-def test_i2v_preprocess_preserves_supported_last_image(last_image) -> None:
-    preprocess = get_wan22_i2v_pre_process_func(SimpleNamespace())
-
-    result = preprocess(_make_i2v_preprocess_request(last_image))
-
-    assert result.prompt["multi_modal_data"]["last_image"] is last_image
-    assert result.batch_compatibility_key == ("wan22_i2v_last_image", True)
-
-
-def test_i2v_preprocess_loads_last_image_path(tmp_path) -> None:
-    preprocess = get_wan22_i2v_pre_process_func(SimpleNamespace())
-    path = tmp_path / "last.png"
-    Image.new("RGB", (16, 16), "blue").save(path)
-
-    result = preprocess(_make_i2v_preprocess_request(str(path)))
-
-    last_image = result.prompt["multi_modal_data"]["last_image"]
-    assert isinstance(last_image, Image.Image)
-    assert last_image.mode == "RGB"
-    assert result.batch_compatibility_key == ("wan22_i2v_last_image", True)
-
-
-def test_i2v_preprocess_rejects_multiple_last_images() -> None:
-    preprocess = get_wan22_i2v_pre_process_func(SimpleNamespace())
-
-    with pytest.raises(ValueError, match="at most one last_image"):
-        preprocess(
-            _make_i2v_preprocess_request(
-                [
-                    Image.new("RGB", (16, 16), "blue"),
-                    Image.new("RGB", (16, 16), "green"),
-                ]
-            )
-        )
-
-
-def test_i2v_preprocess_rejects_unsupported_last_image_type() -> None:
-    preprocess = get_wan22_i2v_pre_process_func(SimpleNamespace())
-
-    with pytest.raises(TypeError, match="Unsupported last_image format"):
-        preprocess(_make_i2v_preprocess_request({"not": "an image"}))
-
-
 def test_i2v_diffuse_selects_stage_guidance_and_expands_timesteps() -> None:
    pipeline = _make_i2v_pipeline(expand_timesteps=True)
    latents = torch.zeros(1, 4, 2, 4, 4)
@@ -244,169 +126,3 @@ def test_i2v_prepare_latents_builds_expand_condition_and_first_frame_mask() -> N
    assert first_frame_mask.shape == (1, 1, 2, 2, 2)
    assert first_frame_mask[:, :, 0].sum() == 0
    assert first_frame_mask[:, :, 1].sum() == 4
-
-
-def test_i2v_prepare_latents_preserves_batched_image_conditions() -> None:
-    pipeline = _make_i2v_pipeline(expand_timesteps=True)
-    generators = [
-        torch.Generator(device="cpu").manual_seed(1),
-        torch.Generator(device="cpu").manual_seed(2),
-    ]
-
-    latents, condition, first_frame_mask = pipeline.prepare_latents(
-        image=torch.zeros(2, 3, 16, 16),
-        batch_size=2,
-        num_channels_latents=4,
-        height=16,
-        width=16,
-        num_frames=5,
-        dtype=torch.float32,
-        device=torch.device("cpu"),
-        generator=generators,
-    )
-
-    assert latents.shape == (2, 4, 2, 2, 2)
-    assert condition.shape == (2, 4, 1, 2, 2)
-    assert first_frame_mask.shape == (2, 1, 2, 2, 2)
-
-
-def test_i2v_forward_batches_conditions_random_inputs_and_outputs(monkeypatch) -> None:
-    monkeypatch.setattr(
-        "vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_i2v.current_omni_platform",
-        SimpleNamespace(is_available=lambda: False),
-    )
-    pipeline = _make_i2v_pipeline(expand_timesteps=True)
-    pipeline.scheduler = StubScheduler([9])
-    pipeline.od_config = SimpleNamespace(flow_shift=5.0)
-    pipeline._sample_solver = "unipc"
-    pipeline._flow_shift = 5.0
-    pipeline.boundary_ratio = 0.875
-    pipeline.has_image_encoder = False
-    pipeline._guidance_scale = None
-    pipeline._guidance_scale_2 = None
-    pipeline._num_timesteps = None
-    pipeline._current_timestep = None
-    pipeline.check_inputs = lambda **kwargs: None
-    prepare_call = {}
-
-    def _fake_encode_prompt(**kwargs):
-        batch_size = len(kwargs["prompt"])
-        n = batch_size * kwargs["num_videos_per_prompt"]
-        return torch.zeros(n, 2, 3), None
-
-    def _fake_prepare_latents(**kwargs):
-        prepare_call.update(kwargs)
-        batch_size = kwargs["batch_size"]
-        return (
-            kwargs["latents"],
-            torch.zeros(batch_size, 4, 1, 2, 2),
-            torch.ones(batch_size, 1, 2, 2, 2),
-        )
-
-    pipeline.encode_prompt = _fake_encode_prompt  # type: ignore[method-assign]
-    pipeline.prepare_latents = _fake_prepare_latents  # type: ignore[method-assign]
-    pipeline.diffuse = lambda **kwargs: kwargs["latents"]  # type: ignore[method-assign]
-
-    gen_a = torch.Generator(device="cpu").manual_seed(1)
-    gen_b = torch.Generator(device="cpu").manual_seed(2)
-    latents_a = torch.zeros(2, 4, 2, 2, 2)
-    latents_b = torch.ones(2, 4, 2, 2, 2)
-    image_a = torch.zeros(1, 3, 16, 16)
-    image_b = torch.ones(1, 3, 16, 16)
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={"prompt": "first", "multi_modal_data": {"image": image_a}},
-                sampling_params=_make_i2v_sampling(
-                    generator=gen_a,
-                    latents=latents_a,
-                    num_outputs_per_prompt=2,
-                ),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={"prompt": "second", "multi_modal_data": {"image": image_b}},
-                sampling_params=_make_i2v_sampling(
-                    generator=gen_b,
-                    latents=latents_b,
-                    num_outputs_per_prompt=2,
-                ),
-            ),
-        ]
-    )
-
-    outputs = pipeline.forward(batch)
-
-    assert prepare_call["batch_size"] == 4
-    assert prepare_call["generator"] == [gen_a, gen_a, gen_b, gen_b]
-    torch.testing.assert_close(prepare_call["latents"], torch.cat([latents_a, latents_b]))
-    torch.testing.assert_close(
-        prepare_call["image"],
-        torch.cat([image_a, image_a, image_b, image_b]),
-    )
-    assert len(outputs) == 2
-    torch.testing.assert_close(outputs[0].output, latents_a)
-    torch.testing.assert_close(outputs[1].output, latents_b)
-
-
-def test_i2v_forward_rejects_mismatched_tensor_condition_shapes(monkeypatch) -> None:
-    monkeypatch.setattr(
-        "vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_i2v.current_omni_platform",
-        SimpleNamespace(is_available=lambda: False),
-    )
-    pipeline = _make_i2v_pipeline(expand_timesteps=True)
-    pipeline.scheduler = StubScheduler([9])
-    pipeline.od_config = SimpleNamespace(flow_shift=5.0)
-    pipeline._sample_solver = "unipc"
-    pipeline._flow_shift = 5.0
-    pipeline.boundary_ratio = 0.875
-    pipeline.has_image_encoder = False
-    pipeline._guidance_scale = None
-    pipeline._guidance_scale_2 = None
-    pipeline._num_timesteps = None
-    pipeline._current_timestep = None
-    pipeline.check_inputs = lambda **kwargs: None
-    pipeline.encode_prompt = lambda **kwargs: (torch.zeros(2, 2, 3), None)  # type: ignore[method-assign]
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={"prompt": "first", "multi_modal_data": {"image": torch.zeros(1, 3, 16, 16)}},
-                sampling_params=_make_i2v_sampling(),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={"prompt": "second", "multi_modal_data": {"image": torch.zeros(1, 3, 16, 8)}},
-                sampling_params=_make_i2v_sampling(),
-            ),
-        ]
-    )
-
-    with pytest.raises(ValueError, match="image condition"):
-        pipeline.forward(batch)
-
-
-def test_i2v_forward_rejects_mixed_last_image_presence() -> None:
-    pipeline = _make_i2v_pipeline(expand_timesteps=True)
-    image = torch.zeros(1, 3, 16, 16)
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={
-                    "prompt": "first",
-                    "multi_modal_data": {"image": image, "last_image": image},
-                },
-                sampling_params=_make_i2v_sampling(),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={"prompt": "second", "multi_modal_data": {"image": image}},
-                sampling_params=_make_i2v_sampling(),
-            ),
-        ]
-    )
-
-    with pytest.raises(ValueError, match="mix of provided and missing last_image"):
-        pipeline.forward(batch)
@@ -93,29 +93,6 @@ def _make_pipeline() -> Wan22Pipeline:
    return pipeline


-def _make_sampling(**overrides):
-    values = {
-        "height": None,
-        "width": None,
-        "num_frames": 1,
-        "num_inference_steps": 2,
-        "guidance_scale_provided": True,
-        "guidance_scale": 1.0,
-        "guidance_scale_2": None,
-        "guidance_scale_2_provided": False,
-        "boundary_ratio": None,
-        "generator": None,
-        "seed": None,
-        "num_outputs_per_prompt": 1,
-        "max_sequence_length": 32,
-        "latents": None,
-        "output_type": "latent",
-        "extra_args": {},
-    }
-    values.update(overrides)
-    return SimpleNamespace(**values)
-
-
@pytest.mark.parametrize(
    ("sampling_params_kwargs", "expected_low", "expected_high"),
    [
@@ -152,10 +129,9 @@ def test_forward_delegates_denoising_to_diffuse(
    )
    batch = DiffusionRequestBatch(requests=[mock_req])

-    outputs = pipeline.forward(batch)
+    output = pipeline.forward(batch)

-    assert len(outputs) == 1
-    assert torch.equal(outputs[0].output, torch.ones((1, 4, 1, 8, 8)))
+    assert torch.equal(output.output, torch.ones((1, 4, 1, 8, 8)))
    assert torch.equal(captured["prompt_embeds"], torch.zeros(1, 32, 8))
    assert torch.equal(captured["timesteps"], pipeline.scheduler.timesteps)
    assert captured["guidance_low"] == expected_low
@@ -166,134 +142,6 @@ def test_forward_delegates_denoising_to_diffuse(
    assert pipeline.scheduler.set_timesteps_calls == [(2, torch.device("cpu"))]


-def test_forward_batches_text_generators_latents_and_splits_outputs() -> None:
-    pipeline = _make_pipeline()
-    encode_call = {}
-    prepare_call = {}
-
-    def _fake_encode_prompt(**kwargs):
-        encode_call.update(kwargs)
-        batch_size = len(kwargs["prompt"])
-        n = batch_size * kwargs["num_videos_per_prompt"]
-        return torch.arange(n, dtype=torch.float32).view(n, 1, 1), torch.zeros(n, 1, 1)
-
-    def _fake_prepare_latents(**kwargs):
-        prepare_call.update(kwargs)
-        return kwargs["latents"]
-
-    pipeline.encode_prompt = _fake_encode_prompt  # type: ignore[method-assign]
-    pipeline.prepare_latents = _fake_prepare_latents  # type: ignore[method-assign]
-    pipeline.diffuse = lambda **kwargs: kwargs["latents"]  # type: ignore[method-assign]
-
-    gen_a = torch.Generator(device="cpu").manual_seed(1)
-    gen_b = torch.Generator(device="cpu").manual_seed(2)
-    latents_a = torch.zeros(2, 4, 1, 2, 2)
-    latents_b = torch.ones(2, 4, 1, 2, 2)
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={"prompt": "first", "negative_prompt": "bad first"},
-                sampling_params=_make_sampling(
-                    generator=gen_a,
-                    latents=latents_a,
-                    num_outputs_per_prompt=2,
-                ),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={"prompt": "second", "negative_prompt": "bad second"},
-                sampling_params=_make_sampling(
-                    generator=gen_b,
-                    latents=latents_b,
-                    num_outputs_per_prompt=2,
-                ),
-            ),
-        ]
-    )
-
-    outputs = pipeline.forward(batch)
-
-    assert encode_call["prompt"] == ["first", "second"]
-    assert encode_call["negative_prompt"] == ["bad first", "bad second"]
-    assert prepare_call["batch_size"] == 4
-    assert prepare_call["generator"] == [gen_a, gen_a, gen_b, gen_b]
-    torch.testing.assert_close(prepare_call["latents"], torch.cat([latents_a, latents_b]))
-    assert len(outputs) == 2
-    torch.testing.assert_close(outputs[0].output, latents_a)
-    torch.testing.assert_close(outputs[1].output, latents_b)
-
-
-def test_forward_batches_precomputed_prompt_embeddings() -> None:
-    pipeline = _make_pipeline()
-    diffuse_call = {}
-    pipeline.encode_prompt = lambda **kwargs: pytest.fail("text encoder must not run")  # type: ignore[method-assign]
-    pipeline.prepare_latents = lambda **kwargs: torch.zeros(kwargs["batch_size"], 4, 1, 2, 2)  # type: ignore[method-assign]
-
-    def _fake_diffuse(**kwargs):
-        diffuse_call.update(kwargs)
-        return kwargs["latents"]
-
-    pipeline.diffuse = _fake_diffuse  # type: ignore[method-assign]
-    embeds_a = torch.zeros(3, 4)
-    embeds_b = torch.ones(3, 4)
-    negative_a = torch.full((3, 4), 2.0)
-    negative_b = torch.full((3, 4), 3.0)
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={"prompt_embeds": embeds_a, "negative_prompt_embeds": negative_a},
-                sampling_params=_make_sampling(guidance_scale=4.0),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={"prompt_embeds": embeds_b, "negative_prompt_embeds": negative_b},
-                sampling_params=_make_sampling(guidance_scale=4.0),
-            ),
-        ]
-    )
-
-    outputs = pipeline.forward(batch)
-
-    torch.testing.assert_close(diffuse_call["prompt_embeds"], torch.stack([embeds_a, embeds_b]))
-    torch.testing.assert_close(diffuse_call["negative_prompt_embeds"], torch.stack([negative_a, negative_b]))
-    assert len(outputs) == 2
-
-
-def test_prepare_latents_with_request_generators_matches_single_generation() -> None:
-    pipeline = _make_pipeline()
-    kwargs = {
-        "num_channels_latents": 4,
-        "height": 16,
-        "width": 16,
-        "num_frames": 5,
-        "dtype": torch.float32,
-        "device": torch.device("cpu"),
-    }
-
-    batched = Wan22Pipeline.prepare_latents(
-        pipeline,
-        batch_size=2,
-        generator=[torch.Generator().manual_seed(1), torch.Generator().manual_seed(2)],
-        **kwargs,
-    )
-    singles = torch.cat(
-        [
-            Wan22Pipeline.prepare_latents(
-                pipeline,
-                batch_size=1,
-                generator=torch.Generator().manual_seed(seed),
-                **kwargs,
-            )
-            for seed in (1, 2)
-        ]
-    )
-
-    torch.testing.assert_close(batched, singles)
-    assert not torch.equal(batched[0], batched[1])
-
-
 def test_diffuse_runs_prediction_and_scheduler_for_each_timestep() -> None:
    pipeline = _make_pipeline()
    latents = torch.zeros((1, 1, 1, 2, 2), dtype=torch.float32)
@@ -10,7 +10,6 @@ import torch
 import vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2 as wan22_module
 from vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2 import (
    create_transformer_from_config,
-    get_wan22_post_process_func,
    load_transformer_config,
    retrieve_latents,
 )
@@ -27,17 +26,6 @@ class _LatentDist:
        return torch.tensor([2.0])


-def test_wan22_postprocess_honors_request_output_type() -> None:
-    video = torch.zeros(1, 4, 1, 2, 2)
-
-    output = get_wan22_post_process_func(SimpleNamespace())(
-        video,
-        sampling_params=SimpleNamespace(output_type="latent"),
-    )
-
-    assert output is video
-
-
 def test_retrieve_latents_supports_sample_mode_argmax_and_direct_latents() -> None:
    generator = torch.Generator(device="cpu")

@@ -1,307 +1,14 @@
-from types import SimpleNamespace
 from unittest.mock import MagicMock, patch

-import numpy as np
-import PIL.Image
 import pytest
 import torch
 import torch.nn as nn

-from vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_s2v import (
-    Wan22S2VPipeline,
-    _make_clip_generators,
-)
 from vllm_omni.diffusion.models.wan2_2.wan2_2_s2v_transformer import WanS2VTransformer3DModel
-from vllm_omni.diffusion.worker.request_batch import DiffusionRequestBatch

 pytestmark = [pytest.mark.core_model, pytest.mark.cpu]


-def _make_s2v_sampling(**overrides):
-    values = {
-        "height": 16,
-        "width": 16,
-        "num_frames": 8,
-        "num_inference_steps": 1,
-        "guidance_scale_provided": True,
-        "guidance_scale": 1.0,
-        "generator": None,
-        "seed": None,
-        "num_outputs_per_prompt": 1,
-        "max_sequence_length": 8,
-        "latents": None,
-        "output_type": "latent",
-        "extra_args": {},
-    }
-    values.update(overrides)
-    return SimpleNamespace(**values)
-
-
-def _make_s2v_validation_pipeline() -> Wan22S2VPipeline:
-    pipeline = object.__new__(Wan22S2VPipeline)
-    nn.Module.__init__(pipeline)
-    pipeline.device = torch.device("cpu")
-    pipeline.transformer = SimpleNamespace(dtype=torch.float32)
-    pipeline.vae_scale_factor_spatial = 8
-    pipeline.resolution_divisor = 16
-    pipeline.motion_frames = 7
-    pipeline.drop_first_motion = True
-    pipeline._DEFAULT_INFER_FRAMES = 8
-    pipeline._guidance_scale = None
-    pipeline.check_inputs = lambda *args, **kwargs: None
-    pipeline.encode_prompt = lambda **kwargs: (torch.zeros(2, 2, 3), None)  # type: ignore[method-assign]
-    return pipeline
-
-
-def test_s2v_predict_noise_keeps_batch_dimension() -> None:
-    pipeline = object.__new__(Wan22S2VPipeline)
-    nn.Module.__init__(pipeline)
-    pipeline.device = torch.device("cpu")
-    transformer = MagicMock()
-    transformer.parameters.return_value = iter([torch.zeros(1, dtype=torch.bfloat16)])
-    transformer.return_value = (torch.zeros(1, 16, 2, 2, 2),)
-    pipeline.transformer = transformer
-
-    result = pipeline.predict_noise(hidden_states=torch.zeros(1, 16, 2, 2, 2))
-
-    assert result.shape == (1, 16, 2, 2, 2)
-
-
-def test_s2v_clip_generators_preserve_main_seed_per_clip_behavior() -> None:
-    explicit_generator = torch.Generator(device="cpu").manual_seed(999)
-
-    generators = _make_clip_generators(
-        seeds=[1234, 5678, None, None],
-        request_generators=[
-            torch.Generator(device="cpu").manual_seed(1234),
-            torch.Generator(device="cpu").manual_seed(5678),
-            explicit_generator,
-            None,
-        ],
-        clip_index=2,
-        device=torch.device("cpu"),
-    )
-
-    assert [generator.initial_seed() for generator in generators if generator is not None] == [1236, 5680, 999, 2]
-    assert generators[2] is explicit_generator
-
-
-def test_s2v_forward_rejects_different_num_repeat() -> None:
-    pipeline = _make_s2v_validation_pipeline()
-    pipeline.encode_audio = lambda *args, **kwargs: (torch.zeros(1, 1, 2, 16), 2, 16)  # type: ignore[method-assign]
-    image = PIL.Image.new("RGB", (16, 16))
-    audio = np.zeros(16000, dtype=np.float32)
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={
-                    "prompt": "first",
-                    "multi_modal_data": {"image": image, "audio": audio},
-                    "additional_information": {"num_repeat": 1},
-                },
-                sampling_params=_make_s2v_sampling(),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={
-                    "prompt": "second",
-                    "multi_modal_data": {"image": image, "audio": audio},
-                    "additional_information": {"num_repeat": 2},
-                },
-                sampling_params=_make_s2v_sampling(),
-            ),
-        ]
-    )
-
-    with pytest.raises(ValueError, match="num_repeat"):
-        pipeline.forward(batch)
-
-
-def test_s2v_forward_rejects_different_audio_time_lengths() -> None:
-    pipeline = _make_s2v_validation_pipeline()
-    pipeline.encode_audio = lambda audio, **kwargs: (  # type: ignore[method-assign]
-        torch.zeros(1, 1, 2, 16 if audio[0] == 0 else 24),
-        1,
-        16 if audio[0] == 0 else 24,
-    )
-    image = PIL.Image.new("RGB", (16, 16))
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={
-                    "prompt": "first",
-                    "multi_modal_data": {"image": image, "audio": np.zeros(16000, dtype=np.float32)},
-                },
-                sampling_params=_make_s2v_sampling(),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={
-                    "prompt": "second",
-                    "multi_modal_data": {"image": image, "audio": np.ones(16000, dtype=np.float32)},
-                },
-                sampling_params=_make_s2v_sampling(),
-            ),
-        ]
-    )
-
-    with pytest.raises(ValueError, match="audio embedding time lengths"):
-        pipeline.forward(batch)
-
-
-def test_s2v_forward_rejects_different_init_first_frame() -> None:
-    pipeline = _make_s2v_validation_pipeline()
-    image = PIL.Image.new("RGB", (16, 16))
-    audio = np.zeros(16000, dtype=np.float32)
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={
-                    "prompt": "first",
-                    "multi_modal_data": {"image": image, "audio": audio},
-                    "additional_information": {"init_first_frame": True},
-                },
-                sampling_params=_make_s2v_sampling(),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={
-                    "prompt": "second",
-                    "multi_modal_data": {"image": image, "audio": audio},
-                    "additional_information": {"init_first_frame": False},
-                },
-                sampling_params=_make_s2v_sampling(),
-            ),
-        ]
-    )
-
-    with pytest.raises(ValueError, match="init_first_frame"):
-        pipeline.forward(batch)
-
-
-def test_s2v_forward_rejects_different_raw_audio_shapes() -> None:
-    pipeline = _make_s2v_validation_pipeline()
-    pipeline.encode_audio = lambda *args, **kwargs: (torch.zeros(1, 1, 2, 16), 1, 16)  # type: ignore[method-assign]
-    image = PIL.Image.new("RGB", (16, 16))
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={
-                    "prompt": "first",
-                    "multi_modal_data": {"image": image, "audio": np.zeros(16000, dtype=np.float32)},
-                },
-                sampling_params=_make_s2v_sampling(),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={
-                    "prompt": "second",
-                    "multi_modal_data": {"image": image, "audio": np.zeros(8000, dtype=np.float32)},
-                },
-                sampling_params=_make_s2v_sampling(),
-            ),
-        ]
-    )
-
-    with pytest.raises(ValueError, match="raw audio shapes and sample rates"):
-        pipeline.forward(batch)
-
-
-def test_s2v_forward_batches_request_local_inputs_and_splits_outputs() -> None:
-    pipeline = object.__new__(Wan22S2VPipeline)
-    nn.Module.__init__(pipeline)
-    pipeline.device = torch.device("cpu")
-    pipeline.transformer = MagicMock()
-    pipeline.transformer.dtype = torch.float32
-    pipeline.transformer.parameters.return_value = iter([torch.zeros(1, dtype=torch.float32)])
-    pipeline.transformer.casual_audio_encoder = None
-    pipeline.transformer.encode_audio.side_effect = lambda audio, _motion: {"audio_emb": audio}
-    pipeline.vae = MagicMock()
-    pipeline.vae.dtype = torch.float32
-    pipeline.vae.decode.return_value = (torch.zeros(4, 3, 8, 16, 16),)
-    pipeline.od_config = SimpleNamespace(
-        enable_cpu_offload=False,
-        parallel_config=SimpleNamespace(use_hsdp=False),
-    )
-    pipeline.scheduler = MagicMock(timesteps=torch.tensor([1.0]))
-    pipeline.vae_scale_factor_spatial = 8
-    pipeline.resolution_divisor = 16
-    pipeline.motion_frames = 7
-    pipeline.drop_first_motion = True
-    pipeline._DEFAULT_INFER_FRAMES = 8
-    pipeline._guidance_scale = None
-    pipeline._num_timesteps = None
-    pipeline.check_inputs = lambda *args, **kwargs: None
-    pipeline.encode_prompt = MagicMock(return_value=(torch.zeros(4, 2, 3), torch.ones(4, 2, 3)))
-    pipeline.encode_audio = MagicMock(return_value=(torch.zeros(1, 1, 2, 8), 1, 8))
-    pipeline.encode_ref_image = MagicMock(return_value=torch.zeros(1, 16, 1, 2, 2))
-    pipeline.prepare_motion_latents = MagicMock(
-        side_effect=lambda pixels, **_: torch.zeros(pixels.shape[0], 16, 2, 2, 2)
-    )
-    pipeline._denormalize_latents = lambda latents: latents
-    pipeline.diffuse = MagicMock(side_effect=lambda **kwargs: kwargs["latents"])
-
-    image = PIL.Image.new("RGB", (16, 16))
-    audio_a = np.zeros(16000, dtype=np.float32)
-    audio_b = np.ones(16000, dtype=np.float32)
-    latents_a = torch.zeros(2, 16, 2, 2, 2)
-    latents_b = torch.ones(2, 16, 2, 2, 2)
-    generators_a = [torch.Generator().manual_seed(1), torch.Generator().manual_seed(2)]
-    generators_b = [torch.Generator().manual_seed(3), torch.Generator().manual_seed(4)]
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={
-                    "prompt": "first",
-                    "negative_prompt": "negative first",
-                    "multi_modal_data": {"image": image, "audio": audio_a},
-                },
-                sampling_params=_make_s2v_sampling(
-                    num_outputs_per_prompt=2,
-                    generator=generators_a,
-                    latents=latents_a,
-                ),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={
-                    "prompt": "second",
-                    "negative_prompt": "negative second",
-                    "multi_modal_data": {"image": image, "audio": audio_b},
-                },
-                sampling_params=_make_s2v_sampling(
-                    num_outputs_per_prompt=2,
-                    generator=generators_b,
-                    latents=latents_b,
-                ),
-            ),
-        ]
-    )
-
-    with patch("vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_s2v.current_omni_platform") as platform:
-        platform.is_available.return_value = False
-        outputs = pipeline.forward(batch)
-
-    assert len(outputs) == 2
-    assert outputs[0].output[0].shape[0] == 2
-    assert outputs[1].output[0].shape[0] == 2
-    assert outputs[0].output[1].shape == (16000,)
-    assert outputs[1].output[1].shape == (16000,)
-    np.testing.assert_array_equal(outputs[0].output[1], audio_a)
-    np.testing.assert_array_equal(outputs[1].output[1], audio_b)
-    assert np.shares_memory(outputs[0].output[1], audio_a)
-    assert np.shares_memory(outputs[1].output[1], audio_b)
-    torch.testing.assert_close(pipeline.diffuse.call_args.kwargs["latents"], torch.cat([latents_a, latents_b]))
-    assert pipeline.diffuse.call_args.kwargs["clip_generator"] == generators_a + generators_b
-    assert pipeline.encode_prompt.call_args.kwargs["prompt"] == ["first", "second"]
-    assert pipeline.encode_prompt.call_args.kwargs["negative_prompt"] == ["negative first", "negative second"]
-
-
 def test_s2v_exposes_hsdp_shard_conditions_for_transformer_blocks():
    model = object.__new__(WanS2VTransformer3DModel)
    nn.Module.__init__(model)
@@ -553,37 +260,31 @@ def test_s2v_pipeline_hsdp_forward_complete_process():
    pipeline._prompt_clean = Wan22S2VPipeline._prompt_clean

    # -- Build request --
-    sampling_params = _make_s2v_sampling(
-        height=704,
-        width=1024,
-        num_frames=80,
-        num_inference_steps=5,
-        guidance_scale=4.5,
-        generator=torch.Generator(device="cpu").manual_seed(42),
-        max_sequence_length=512,
-    )
+    from vllm_omni.diffusion.request import OmniDiffusionRequest
+
+    sampling_params = MagicMock()
+    sampling_params.height = 704
+    sampling_params.width = 1024
+    sampling_params.num_frames = 80
+    sampling_params.num_inference_steps = 5
+    sampling_params.guidance_scale = 4.5
+    sampling_params.guidance_scale_provided = True
+    sampling_params.generator = None
+    sampling_params.seed = 42

    ref_image = PIL.Image.new("RGB", (1024, 704))
    audio_data = np.zeros(16000, dtype=np.float32)

-    req = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="test",
-                prompt={
-                    "prompt": "test prompt",
-                    "negative_prompt": "bad quality",
-                    "multi_modal_data": {"image": ref_image, "audio": audio_data},
-                    "additional_information": {
-                        "audio_path": audio_data,
-                        "pose_video": None,
-                        "init_first_frame": False,
-                    },
-                },
-                sampling_params=sampling_params,
-            )
-        ]
-    )
+    req = MagicMock(spec=OmniDiffusionRequest)
+    req.prompts = [
+        {
+            "prompt": "test prompt",
+            "negative_prompt": "bad quality",
+            "multi_modal_data": {"image": ref_image, "audio": audio_data},
+            "additional_information": {"audio_path": audio_data, "pose_video": None, "init_first_frame": False},
+        }
+    ]
+    req.sampling_params = sampling_params

    # -- Mock methods that use complex internal state --
    pipeline.encode_audio = MagicMock(return_value=(torch.zeros(1, 25, 64, 80), 1, 80))
@@ -594,7 +295,7 @@ def test_s2v_pipeline_hsdp_forward_complete_process():
    pipeline.progress_bar.return_value.__exit__ = MagicMock(return_value=False)

    # Mock predict_noise_maybe_with_cfg from CFGParallelMixin
-    pipeline.predict_noise_maybe_with_cfg = MagicMock(return_value=torch.zeros(1, 16, 20, 88, 128))
+    pipeline.predict_noise_maybe_with_cfg = MagicMock(return_value=torch.zeros(16, 20, 88, 128))
    pipeline.predict_noise = Wan22S2VPipeline.predict_noise.__get__(pipeline)

    # Mock platform methods
@@ -605,7 +306,7 @@ def test_s2v_pipeline_hsdp_forward_complete_process():
        with patch(
            "vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_s2v.load_audio", return_value=(audio_data, 16000)
        ):
-            result = Wan22S2VPipeline.forward(pipeline, req=req)[0]
+            result = Wan22S2VPipeline.forward(pipeline, req=req)

    # -- Assertions --
    # Text encoder was called
@@ -8,14 +8,13 @@ import torch
 from PIL import Image
 from torch import nn

-from tests.diffusion.models.wan2_2.conftest import StubScheduler, StubTransformer, StubVAE, noop_progress_bar
+from tests.diffusion.models.wan2_2.conftest import StubTransformer, StubVAE, noop_progress_bar
 from vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_vace import (
    Wan22VACEPipeline,
    create_vace_transformer_from_config,
    get_wan22_vace_pre_process_func,
 )
 from vllm_omni.diffusion.models.wan2_2.wan2_2_vace_transformer import WanVACETransformer3DModel
-from vllm_omni.diffusion.worker.request_batch import DiffusionRequestBatch

 pytestmark = [pytest.mark.core_model, pytest.mark.cpu, pytest.mark.diffusion]

@@ -33,27 +32,6 @@ def _make_vace_pipeline() -> Wan22VACEPipeline:
    return pipeline


-def _make_vace_sampling(**overrides):
-    values = {
-        "height": 16,
-        "width": 16,
-        "num_frames": 5,
-        "num_inference_steps": 1,
-        "guidance_scale_provided": True,
-        "guidance_scale": 1.0,
-        "boundary_ratio": None,
-        "generator": None,
-        "seed": None,
-        "num_outputs_per_prompt": 1,
-        "max_sequence_length": 8,
-        "latents": None,
-        "output_type": "latent",
-        "extra_args": {},
-    }
-    values.update(overrides)
-    return SimpleNamespace(**values)
-
-
 def test_vace_preprocess_collects_reference_video_and_mask_inputs() -> None:
    preprocess = get_wan22_vace_pre_process_func(SimpleNamespace())
    ref = Image.new("RGB", (320, 160), "green")
@@ -160,189 +138,3 @@ def test_vace_diffuse_passes_context_and_scale_to_cfg_branches() -> None:
    assert calls[0]["positive_kwargs"]["vace_context"] is vace_context
    assert calls[0]["negative_kwargs"]["vace_context_scale"] == 0.75
    torch.testing.assert_close(result, torch.ones_like(latents))
-
-
-def test_vace_forward_batches_random_inputs_and_splits_outputs(monkeypatch) -> None:
-    monkeypatch.setattr(
-        "vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_vace.current_omni_platform",
-        SimpleNamespace(is_available=lambda: False),
-    )
-    pipeline = _make_vace_pipeline()
-    pipeline.transformer.vace_patch_embedding = None
-    pipeline.transformer_2 = None
-    pipeline.scheduler = StubScheduler([9])
-    pipeline.od_config = SimpleNamespace(flow_shift=3.0)
-    pipeline._sample_solver = "unipc"
-    pipeline._flow_shift = 3.0
-    pipeline.boundary_ratio = None
-    pipeline._guidance_scale = None
-    pipeline._num_timesteps = None
-    pipeline._current_timestep = None
-    pipeline.check_inputs = lambda **kwargs: None
-    prepare_call = {}
-
-    def _fake_encode_prompt(**kwargs):
-        batch_size = len(kwargs["prompt"])
-        n = batch_size * kwargs["num_videos_per_prompt"]
-        return torch.zeros(n, 2, 3), None
-
-    def _fake_prepare_latents(**kwargs):
-        prepare_call.update(kwargs)
-        return kwargs["latents"]
-
-    pipeline.encode_prompt = _fake_encode_prompt  # type: ignore[method-assign]
-    pipeline.prepare_latents = _fake_prepare_latents  # type: ignore[method-assign]
-    pipeline.diffuse = lambda **kwargs: kwargs["latents"]  # type: ignore[method-assign]
-    gen_a = torch.Generator(device="cpu").manual_seed(1)
-    gen_b = torch.Generator(device="cpu").manual_seed(2)
-    latents_a = torch.zeros(2, 4, 2, 2, 2)
-    latents_b = torch.ones(2, 4, 2, 2, 2)
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={"prompt": "first"},
-                sampling_params=_make_vace_sampling(
-                    generator=gen_a,
-                    latents=latents_a,
-                    num_outputs_per_prompt=2,
-                ),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={"prompt": "second"},
-                sampling_params=_make_vace_sampling(
-                    generator=gen_b,
-                    latents=latents_b,
-                    num_outputs_per_prompt=2,
-                ),
-            ),
-        ]
-    )
-
-    outputs = pipeline.forward(batch)
-
-    assert prepare_call["batch_size"] == 4
-    assert prepare_call["generator"] == [gen_a, gen_a, gen_b, gen_b]
-    torch.testing.assert_close(prepare_call["latents"], torch.cat([latents_a, latents_b]))
-    assert len(outputs) == 2
-    torch.testing.assert_close(outputs[0].output, latents_a)
-    torch.testing.assert_close(outputs[1].output, latents_b)
-
-
-def test_vace_forward_rejects_mismatched_condition_shapes(monkeypatch) -> None:
-    monkeypatch.setattr(
-        "vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_vace.current_omni_platform",
-        SimpleNamespace(is_available=lambda: False),
-    )
-    pipeline = _make_vace_pipeline()
-    pipeline.transformer.vace_patch_embedding = object()
-    pipeline.transformer_2 = None
-    pipeline.scheduler = StubScheduler([9])
-    pipeline.od_config = SimpleNamespace(flow_shift=3.0)
-    pipeline._sample_solver = "unipc"
-    pipeline._flow_shift = 3.0
-    pipeline.boundary_ratio = None
-    pipeline._guidance_scale = None
-    pipeline._num_timesteps = None
-    pipeline._current_timestep = None
-    pipeline.check_inputs = lambda **kwargs: None
-    pipeline.encode_prompt = lambda **kwargs: (torch.zeros(2, 2, 3), None)  # type: ignore[method-assign]
-
-    def _fake_preprocess(video, **kwargs):
-        del kwargs
-        mask = torch.ones_like(video)
-        return video, mask, [[]]
-
-    pipeline.preprocess_conditions = _fake_preprocess  # type: ignore[method-assign]
-    pipeline.prepare_video_latents = (  # type: ignore[method-assign]
-        lambda video, mask, refs, generator, device: torch.zeros(1, 8, video.shape[2], 2, 2)
-    )
-    pipeline.prepare_masks = lambda mask, refs: torch.zeros(1, 4, mask.shape[2], 2, 2)  # type: ignore[method-assign]
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={
-                    "prompt": "first",
-                    "additional_information": {"source_video": torch.zeros(1, 3, 5, 16, 16)},
-                },
-                sampling_params=_make_vace_sampling(),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={
-                    "prompt": "second",
-                    "additional_information": {"source_video": torch.zeros(1, 3, 3, 16, 16)},
-                },
-                sampling_params=_make_vace_sampling(),
-            ),
-        ]
-    )
-
-    with pytest.raises(ValueError, match="VACE condition"):
-        pipeline.forward(batch)
-
-
-def test_vace_forward_rejects_mismatched_reference_image_counts(monkeypatch) -> None:
-    monkeypatch.setattr(
-        "vllm_omni.diffusion.models.wan2_2.pipeline_wan2_2_vace.current_omni_platform",
-        SimpleNamespace(is_available=lambda: False),
-    )
-    pipeline = _make_vace_pipeline()
-    pipeline.transformer.vace_patch_embedding = object()
-    pipeline.transformer_2 = None
-    pipeline.scheduler = StubScheduler([9])
-    pipeline.od_config = SimpleNamespace(flow_shift=3.0)
-    pipeline._sample_solver = "unipc"
-    pipeline._flow_shift = 3.0
-    pipeline.boundary_ratio = None
-    pipeline._guidance_scale = None
-    pipeline._num_timesteps = None
-    pipeline._current_timestep = None
-    pipeline.check_inputs = lambda **kwargs: None
-    pipeline.encode_prompt = lambda **kwargs: (torch.zeros(2, 2, 3), None)  # type: ignore[method-assign]
-
-    def _fake_preprocess(video, mask, reference_images, **kwargs):
-        del kwargs
-        return video, mask, [reference_images]
-
-    pipeline.preprocess_conditions = _fake_preprocess  # type: ignore[method-assign]
-    pipeline.prepare_video_latents = (  # type: ignore[method-assign]
-        lambda video, mask, refs, generator, device: torch.zeros(1, 8, 5, 2, 2)
-    )
-    pipeline.prepare_masks = lambda mask, refs: torch.zeros(1, 4, 5, 2, 2)  # type: ignore[method-assign]
-    video = torch.zeros(1, 3, 5, 16, 16)
-    mask = torch.ones_like(video)
-    reference = torch.zeros(3, 16, 16)
-    batch = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="a",
-                prompt={
-                    "prompt": "first",
-                    "additional_information": {
-                        "source_video": video,
-                        "mask": mask,
-                        "reference_images": [reference],
-                    },
-                },
-                sampling_params=_make_vace_sampling(),
-            ),
-            SimpleNamespace(
-                request_id="b",
-                prompt={
-                    "prompt": "second",
-                    "additional_information": {
-                        "source_video": video,
-                        "mask": mask,
-                        "reference_images": [reference, reference],
-                    },
-                },
-                sampling_params=_make_vace_sampling(),
-            ),
-        ]
-    )
-
-    with pytest.raises(ValueError, match="same number of reference images"):
-        pipeline.forward(batch)
@@ -18,11 +18,6 @@ from torch import nn
 from torch.distributed.tensor import DeviceMesh, DTensor, Replicate

 import vllm_omni.diffusion.offloader.distributed_layerwise_backend as dist_backend_module
-from vllm_omni.diffusion.model_loader.host_weight_plan import (
-    HostWeightPlan,
-    TensorBinding,
-    build_checkpoint_mmap_plan,
-)
 from vllm_omni.diffusion.offloader.base import OffloadConfig, OffloadStrategy
 from vllm_omni.diffusion.offloader.block_discovery import (
    get_blocks_attr_names,
@@ -35,10 +30,7 @@ from vllm_omni.diffusion.offloader.distributed_layerwise_backend import (
    PinnedResidentLayerGroup,
 )
 from vllm_omni.diffusion.offloader.module_residency import PinnedModuleStager
-from vllm_omni.diffusion.offloader.offload_plan import (
-    OffloadPlan,
-    get_offload_plan,
-)
+from vllm_omni.diffusion.offloader.offload_plan import OffloadPlan, get_offload_plan
 from vllm_omni.platforms import current_omni_platform

 pytestmark = [pytest.mark.diffusion, pytest.mark.cpu, pytest.mark.core_model]
@@ -56,9 +48,6 @@ class DummyEvent:
    def record(self, _stream) -> None:
        return None

-    def synchronize(self) -> None:
-        return None
-

@contextmanager
 def dummy_stream(_stream):
@@ -150,41 +139,6 @@ class TestDistributedLayerwiseOffloadHook:
        assert shard.numel() == 4
        assert torch.equal(shard, _make_values(10.0))

-    def test_prefetch_preserves_transposed_weight_stride(
-        self,
-        dist_group,
-        patched_offload_runtime,
-    ):
-        """Online-FP8 Cutlass weights must retain their transposed layout."""
-
-        class StridedBlock(nn.Module):
-            def __init__(self, start: float):
-                super().__init__()
-                base = torch.arange(start, start + 12).reshape(3, 4)
-                self.weight = nn.Parameter(base.t(), requires_grad=False)
-
-        current_block = StridedBlock(1.0)
-        next_block = StridedBlock(20.0)
-        expected = next_block.weight.detach().clone()
-        expected_stride = next_block.weight.stride()
-        assert expected_stride == (1, 4)
-
-        hook = DistributedLayerwiseOffloadHook(
-            next_block=next_block,
-            device=torch.device("cpu"),
-            dp_group=None,
-            dp_size=1,
-            rank=0,
-            copy_stream=DummyStream(),
-            comm_stream=DummyStream(),
-            pin_memory=False,
-        )
-        hook.initialize_hook(current_block)
-        hook.prefetch_layer(slot=0, non_blocking=False)
-
-        assert next_block.weight.stride() == expected_stride
-        assert torch.equal(next_block.weight, expected)
-
    def test_dtensor_wrapper_preserved_across_prefetch_and_offload(self, dist_group, patched_offload_runtime):
        """DTensor wrapper should be preserved through prefetch/offload cycle."""
        current_block = TinyBlock(_make_values(1.0))
@@ -362,9 +316,8 @@ class TestDistributedLayerwiseOffloadHook:
        block.weight = nn.Parameter(torch.zeros(4, dtype=torch.float32))
        next_block = nn.Module()
        next_block.weight = nn.Parameter(torch.arange(4, dtype=torch.float32))
-
-        def transform(tensor):
-            return tensor.flip(0)
+        next_block.weight.mmap_weight_transform = lambda tensor: tensor.flip(0)
+        next_block.weight.mmap_weight_transform_pending = True

        hook = DistributedLayerwiseOffloadHook(
            next_block=next_block,
@@ -373,17 +326,17 @@ class TestDistributedLayerwiseOffloadHook:
            dp_size=2,
            rank=0,
            pin_memory=False,
-            tensor_transforms={id(next_block.weight): transform},
        )
        hook.initialize_hook(block)

        assert torch.equal(hook.cpu_shards[torch.float32], torch.tensor([3.0, 2.0]))

-    def test_transform_is_not_applied_without_a_loader_plan(self, dist_group, patched_offload_runtime):
+    def test_regular_loader_weight_does_not_reapply_mmap_transform(self, dist_group, patched_offload_runtime):
        block = nn.Module()
        block.weight = nn.Parameter(torch.zeros(4, dtype=torch.float32))
        next_block = nn.Module()
        next_block.weight = nn.Parameter(torch.arange(4, dtype=torch.float32))
+        next_block.weight.mmap_weight_transform = lambda tensor: tensor.flip(0)

        hook = DistributedLayerwiseOffloadHook(
            next_block=next_block,
@@ -397,82 +350,6 @@ class TestDistributedLayerwiseOffloadHook:

        assert torch.equal(hook.cpu_shards[torch.float32], torch.tensor([0.0, 1.0]))

-    def test_rank_local_mmap_retains_source_and_uses_staging(self, dist_group, patched_offload_runtime):
-        current_block = nn.Linear(2, 2, bias=False)
-        next_block = nn.Linear(2, 2, bias=False)
-        next_block.weight.data.copy_(torch.arange(4, dtype=torch.float32).view(2, 2))
-        source_storage = next_block.weight.untyped_storage().data_ptr()
-
-        def transform(tensor):
-            return tensor.t()
-
-        hook = DistributedLayerwiseOffloadHook(
-            next_block=next_block,
-            device=torch.device("cpu"),
-            dp_group=None,
-            dp_size=1,
-            rank=0,
-            pin_memory=False,
-            rank_local_mmap=True,
-            tensor_transforms={id(next_block.weight): transform},
-        )
-        hook.initialize_hook(current_block)
-
-        source = hook.cpu_sources[torch.float32][0]["tensor"]
-        assert source.untyped_storage().data_ptr() == source_storage
-        assert hook.cpu_shards == {}
-        assert next_block.weight.numel() == 0
-
-        hook.cpu_staging_buffers = [
-            {torch.float32: torch.empty(4)},
-            {torch.float32: torch.empty(4)},
-        ]
-        hook.prefetch_layer(slot=0, non_blocking=False)
-
-        assert torch.equal(
-            next_block.weight,
-            torch.tensor([[0.0, 2.0], [1.0, 3.0]]),
-        )
-        assert next_block.weight.stride() == (1, 2)
-        assert torch.equal(source, torch.arange(4, dtype=torch.float32).view(2, 2))
-
-    def test_rank_local_mmap_staging_is_bounded_by_largest_block(self, patched_offload_runtime):
-        hooks = []
-        for size in (4, 9):
-            current_block = nn.Linear(1, 1, bias=False)
-            next_block = nn.Module()
-            next_block.weight = nn.Parameter(torch.arange(size, dtype=torch.float32))
-            hook = DistributedLayerwiseOffloadHook(
-                next_block=next_block,
-                device=torch.device("cpu"),
-                dp_group=None,
-                dp_size=1,
-                rank=0,
-                pin_memory=False,
-                shared_buffers=[None, None],
-                rank_local_mmap=True,
-            )
-            hook.initialize_hook(current_block)
-            hooks.append(hook)
-
-        resident_block = nn.Module()
-        resident_block.weight = nn.Parameter(torch.arange(12, dtype=torch.float32))
-        resident_group = PinnedResidentLayerGroup(
-            [resident_block],
-            device=torch.device("cpu"),
-            copy_stream=DummyStream(),
-            pin_memory=False,
-            rank_local_mmap=True,
-        )
-        staging = DistributedLayerwiseOffloadBackend._allocate_shared_cpu_staging_buffers(
-            hooks,
-            resident_group,
-        )
-
-        assert len(staging) == 2
-        assert staging[0][torch.float32].numel() == 12
-        assert staging[1][torch.float32].numel() == 12
-

 class TestPinnedResidentLayerGroup:
    def test_load_offload_reuses_pinned_master_weights(self, patched_offload_runtime):
@@ -509,59 +386,23 @@ class TestPinnedResidentLayerGroup:

        group.offload()

-    def test_load_preserves_noncontiguous_weight_stride(self, patched_offload_runtime):
-        block = nn.Module()
-        weight = torch.arange(6, dtype=torch.float32).reshape(2, 3).t()
-        expected = weight.detach().clone()
-        expected_stride = weight.stride()
-        block.weight = nn.Parameter(weight)
-        group = PinnedResidentLayerGroup(
-            [block],
-            device=torch.device("cpu"),
-            copy_stream=DummyStream(),
-            pin_memory=False,
-        )
+    def test_mmap_loader_attrs_survive_to_empty_parameter_replacement(self):
+        module = nn.Linear(2, 2, bias=False)

-        group.load()
-        assert torch.equal(block.weight, expected)
-        assert block.weight.stride() == expected_stride
-        group.offload()
+        def transform(tensor):
+            return tensor.flip(0)

-    def test_rank_local_mmap_resident_layers_retain_sources(self, patched_offload_runtime):
-        blocks = [nn.Linear(2, 2, bias=False) for _ in range(3)]
-        expected = []
-        source_storage = []
-        for index, block in enumerate(blocks):
-            block.weight.data.fill_(index + 1)
-            expected.append(block.weight.detach().clone())
-            source_storage.append(block.weight.untyped_storage().data_ptr())
+        module.weight.mmap_weight_transform = transform

-        group = PinnedResidentLayerGroup(
-            blocks,
-            device=torch.device("cpu"),
-            copy_stream=DummyStream(),
-            pin_memory=False,
-            rank_local_mmap=True,
-        )
+        backend = object.__new__(DistributedLayerwiseOffloadBackend)
+        backend._remember_mmap_param_attrs(module)
+        module.to_empty(device="meta")

-        for index, state in enumerate(group._states):
-            source = state["cpu_sources"][torch.float32][0]["tensor"]
-            assert source.untyped_storage().data_ptr() == source_storage[index]
-            assert state["cpu_shards"] == {}
-        assert len(group._cpu_staging_buffers) == 2
+        replacement = nn.Parameter(torch.arange(4).reshape(2, 2).float())
+        backend._attach_mmap_param_attrs("weight", replacement, module.weight)

-        group.load()
-        for block, weight in zip(blocks, expected):
-            assert torch.equal(block.weight, weight)
-
-        blocks[0].weight.data.zero_()
-        group.offload()
-        group.load()
-        for block, weight in zip(blocks, expected):
-            assert torch.equal(block.weight, weight)
-
-        group.offload()
-        assert all(block.weight.numel() == 0 for block in blocks)
+        assert replacement.mmap_weight_transform is transform
+        assert replacement.mmap_weight_transform_pending is True

    def test_shared_allgather_output_is_narrowed_to_current_block(
        self, dist_group, patched_offload_runtime, monkeypatch
@@ -704,6 +545,7 @@ class TestMmapWeightLoading:
            OffloadConfig(
                strategy=OffloadStrategy.DISTRIBUTED_LAYER_WISE,
                pin_cpu_memory=False,
+                model_path=str(tmp_path),
            ),
            torch.device("cpu"),
        )
@@ -711,81 +553,13 @@ class TestMmapWeightLoading:
            dits=[pipeline.transformer],
            dit_names=["transformer"],
        )
-        plan = HostWeightPlan(
-            backing_kind="checkpoint_mmap",
-            bindings={
-                name: TensorBinding(
-                    checkpoint_key=name,
-                    file_path=str(weight_file),
-                )
-                for name in weights
-            },
-        )

-        backend._load_weights_via_mmap(pipeline, modules, plan)
+        backend._load_weights_via_mmap(pipeline, modules)

        assert pipeline.transformer.post_load_calls == 1
        assert pipeline.transformer.time_embedder.weight.dtype == torch.float32
        assert pipeline.transformer.blocks[0].weight.dtype == torch.bfloat16

-    def test_allgather_flag_with_group_size_one_uses_rank_local_mmap(
-        self,
-        tmp_path,
-        patched_offload_runtime,
-    ):
-        class Transformer(nn.Module):
-            _layerwise_offload_blocks_attrs = ["blocks"]
-
-            def __init__(self):
-                super().__init__()
-                self.blocks = nn.ModuleList(
-                    [
-                        nn.Linear(2, 2, bias=False),
-                        nn.Linear(2, 2, bias=False),
-                    ]
-                )
-
-        class Pipeline(nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.transformer = Transformer()
-
-            @staticmethod
-            def _remap_ckpt_key(key):
-                return key
-
-        pipeline = Pipeline()
-        weights = {name: torch.ones_like(param) for name, param in pipeline.named_parameters()}
-        checkpoint_file = tmp_path / "model.safetensors"
-        save_file(weights, str(checkpoint_file))
-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=(),
-            model_path=str(tmp_path),
-            tensor_parallel_size=1,
-            use_hsdp=False,
-            online_quantization=False,
-        )
-        assert result.plan is not None
-        backend = DistributedLayerwiseOffloadBackend(
-            OffloadConfig(
-                strategy=OffloadStrategy.DISTRIBUTED_LAYER_WISE,
-                pin_cpu_memory=False,
-                dp_size=1,
-                dlo_use_allgather=True,
-            ),
-            torch.device("cpu"),
-            host_weight_plan=result.plan,
-        )
-
-        backend.enable(pipeline)
-
-        assert backend._using_rank_local_mmap
-        assert backend.dp_group is None
-        assert all(hook.rank_local_mmap for group in backend._all_hook_groups for hook in group)
-        backend.disable()
-

 class TestGetBlocksFromDit:
    def test_get_blocks_from_dit_single_block_attr(self):
@@ -1051,123 +825,23 @@ class TestOffloadPlan:


 class TestMmapValidation:
-    """Tests for loader preflight and backend plan realization."""
+    """Tests for mmap loader validation: TP rejection, strict check, validate_loaded_weights."""

-    def test_component_source_prefix_builds_model_namespace(self, tmp_path):
-        source_root = tmp_path / "partition"
-        checkpoint_dir = source_root / "transformer"
-        checkpoint_dir.mkdir(parents=True)
-        checkpoint_file = checkpoint_dir / "model.safetensors"
-        save_file(
-            {"blocks.0.weight": torch.arange(4, dtype=torch.float32).reshape(1, 4)},
-            str(checkpoint_file),
+    def test_tp_rejected_when_tp_world_size_gt_1(self, tmp_path, patched_offload_runtime, monkeypatch):
+        """DLO+AllGather should reject when TP world size > 1."""
+        import json
+        from types import SimpleNamespace
+
+        from safetensors.torch import save_file
+
+        from vllm_omni.diffusion.offloader.base import OffloadConfig, OffloadStrategy
+
+        # Mock TP world size = 2
+        monkeypatch.setattr(
+            "vllm.distributed.parallel_state.get_tensor_model_parallel_world_size",
+            lambda: 2,
        )

-        class Pipeline(nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.transformer = nn.Module()
-                self.transformer.blocks = nn.ModuleList([nn.Linear(4, 1, bias=False)])
-                self.weights_sources = [
-                    SimpleNamespace(
-                        model_or_path=str(source_root),
-                        subfolder="transformer",
-                        revision=None,
-                        prefix="transformer.",
-                    )
-                ]
-
-        pipeline = Pipeline()
-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=tuple(pipeline.weights_sources),
-            model_path=None,
-            tensor_parallel_size=1,
-            use_hsdp=False,
-            online_quantization=False,
-        )
-
-        assert result.fallback_reason is None
-        assert result.plan is not None
-        assert result.plan.bindings["transformer.blocks.0.weight"] == TensorBinding(
-            checkpoint_key="blocks.0.weight",
-            file_path=str(checkpoint_file),
-        )
-        assert result.plan.planned_source_prefixes == frozenset({"transformer."})
-
-    def test_non_dedicated_component_source_falls_back_before_discovery(self):
-        pipeline = nn.Module()
-        pipeline.transformer = nn.Linear(2, 2, bias=False)
-        sources = (
-            SimpleNamespace(
-                model_or_path="unused",
-                subfolder=None,
-                revision=None,
-                prefix="",
-            ),
-        )
-
-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=sources,
-            model_path=None,
-            tensor_parallel_size=1,
-            use_hsdp=False,
-            online_quantization=False,
-        )
-
-        assert result.plan is None
-        assert "dedicated component weight source" in result.fallback_reason
-
-    def test_preflight_falls_back_before_mutation_for_missing_model_tensors(self, tmp_path):
-        source_root = tmp_path / "partition"
-        checkpoint_dir = source_root / "transformer"
-        checkpoint_dir.mkdir(parents=True)
-        save_file(
-            {"blocks.0.weight": torch.ones((2, 2), dtype=torch.float32)},
-            str(checkpoint_dir / "model.safetensors"),
-        )
-
-        class Transformer(nn.Module):
-            _layerwise_offload_blocks_attrs = ["blocks"]
-
-            def __init__(self):
-                super().__init__()
-                self.blocks = nn.ModuleList([nn.Linear(2, 2, bias=False), nn.Linear(2, 2, bias=False)])
-
-        class Pipeline(nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.transformer = Transformer()
-                self.weights_sources = [
-                    SimpleNamespace(
-                        model_or_path=str(source_root),
-                        subfolder="transformer",
-                        revision=None,
-                        prefix="transformer.",
-                    )
-                ]
-
-        pipeline = Pipeline()
-        original_weight = pipeline.transformer.blocks[0].weight
-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=tuple(pipeline.weights_sources),
-            model_path=None,
-            tensor_parallel_size=1,
-            use_hsdp=False,
-            online_quantization=False,
-        )
-
-        assert result.plan is None
-        assert "1 required DiT tensors" in result.fallback_reason
-        assert pipeline.transformer.blocks[0].weight is original_weight
-        assert not original_weight.is_meta
-
-    def test_tp_falls_back_before_checkpoint_discovery(self):
        class SimpleModel(nn.Module):
            _layerwise_offload_blocks_attrs = ["blocks"]

@@ -1185,118 +859,41 @@ class TestMmapValidation:
                return key

        pipeline = SimplePipeline()
-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=(),
-            model_path="unused",
-            tensor_parallel_size=2,
-            use_hsdp=False,
-            online_quantization=False,
+        weights = {name: torch.ones(p.shape, dtype=p.dtype) for name, p in pipeline.named_parameters() if not p.is_meta}
+        save_file(weights, str(tmp_path / "model.safetensors"))
+        (tmp_path / "model.safetensors.index.json").write_text(
+            json.dumps({"weight_map": {k: "model.safetensors" for k in weights}})
        )

-        assert result.plan is None
-        assert result.fallback_reason == "TP=2 requires the ordinary loader"
+        backend = DistributedLayerwiseOffloadBackend(
+            OffloadConfig(
+                strategy=OffloadStrategy.DISTRIBUTED_LAYER_WISE,
+                pin_cpu_memory=False,
+                model_path=str(tmp_path),
+            ),
+            torch.device("cpu"),
+        )
+        modules = SimpleNamespace(dits=[pipeline.transformer], dit_names=["transformer"])

-    @pytest.mark.parametrize(
-        ("use_hsdp", "online_quantization", "expected_reason"),
-        [
-            (True, False, "HSDP requires the ordinary loader"),
-            (False, True, "online quantization requires the ordinary loader"),
-        ],
-    )
-    def test_runtime_layouts_that_need_loader_callbacks_fall_back(
-        self,
-        use_hsdp,
-        online_quantization,
-        expected_reason,
-    ):
-        pipeline = nn.Module()
-        pipeline.transformer = nn.Linear(2, 2, bias=False)
+        with pytest.raises(ValueError, match="Tensor Parallel"):
+            backend._load_weights_via_mmap(pipeline, modules)

-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=(),
-            model_path="unused",
-            tensor_parallel_size=1,
-            use_hsdp=use_hsdp,
-            online_quantization=online_quantization,
+    def test_tp_allowed_when_tp_world_size_eq_1(self, tmp_path, patched_offload_runtime, monkeypatch):
+        """DLO+AllGather should work when TP world size = 1, even if params
+        have custom weight_loader attributes (e.g. QKVParallelLinear at TP=1)."""
+        import json
+        from types import SimpleNamespace
+
+        from safetensors.torch import save_file
+
+        from vllm_omni.diffusion.offloader.base import OffloadConfig, OffloadStrategy
+
+        # Mock TP world size = 1 (default)
+        monkeypatch.setattr(
+            "vllm.distributed.parallel_state.get_tensor_model_parallel_world_size",
+            lambda: 1,
        )

-        assert result.plan is None
-        assert result.fallback_reason == expected_reason
-
-    @pytest.mark.parametrize(
-        ("checkpoint_tensor", "expected_reason"),
-        [
-            (torch.ones((3, 2), dtype=torch.float32), "shape mismatch"),
-            (torch.ones((2, 2), dtype=torch.bfloat16), "dtype mismatch"),
-        ],
-    )
-    def test_source_metadata_mismatch_falls_back(self, tmp_path, checkpoint_tensor, expected_reason):
-        checkpoint_file = tmp_path / "model.safetensors"
-        save_file({"transformer.weight": checkpoint_tensor}, str(checkpoint_file))
-
-        class Pipeline(nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.transformer = nn.Linear(2, 2, bias=False)
-
-            @staticmethod
-            def _remap_ckpt_key(key):
-                return key
-
-        pipeline = Pipeline()
-
-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=(),
-            model_path=str(tmp_path),
-            tensor_parallel_size=1,
-            use_hsdp=False,
-            online_quantization=False,
-        )
-
-        assert result.plan is None
-        assert expected_reason in result.fallback_reason
-
-    def test_non_persistent_buffers_do_not_need_checkpoint_bindings(self, tmp_path):
-        class Transformer(nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.weight = nn.Parameter(torch.ones(2, 2))
-                self.register_buffer("derived", torch.ones(2), persistent=False)
-
-        checkpoint_file = tmp_path / "model.safetensors"
-        save_file({"transformer.weight": torch.ones(2, 2)}, str(checkpoint_file))
-
-        class Pipeline(nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.transformer = Transformer()
-
-            @staticmethod
-            def _remap_ckpt_key(key):
-                return key
-
-        pipeline = Pipeline()
-
-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=(),
-            model_path=str(tmp_path),
-            tensor_parallel_size=1,
-            use_hsdp=False,
-            online_quantization=False,
-        )
-
-        assert result.plan is not None
-        assert set(result.plan.bindings) == {"transformer.weight"}
-
-    def test_unadapted_custom_weight_loader_falls_back_at_tp1(self, tmp_path):
        def custom_weight_loader(param, weight):
            param.data.copy_(weight)

@@ -1326,67 +923,27 @@ class TestMmapValidation:
            json.dumps({"weight_map": {k: "model.safetensors" for k in weights}})
        )

-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=(),
-            model_path=str(tmp_path),
-            tensor_parallel_size=1,
-            use_hsdp=False,
-            online_quantization=False,
-        )
-
-        assert result.plan is None
-        assert "custom weight_loader" in result.fallback_reason
-
-    def test_cosmos3_adapter_preserves_existing_tp1_direct_mmap_contract(self, tmp_path):
-        from vllm_omni.diffusion.models.cosmos3.pipeline_cosmos3 import (
-            Cosmos3OmniDiffusersPipeline,
-        )
-        from vllm_omni.diffusion.models.cosmos3.transformer_cosmos3 import (
-            Cosmos3VFMTransformer,
-        )
-
-        transformer = object.__new__(Cosmos3VFMTransformer)
-        nn.Module.__init__(transformer)
-        transformer.proj_in = nn.Linear(2, 2, bias=False)
-        transformer.proj_in.weight.weight_loader = lambda *_args: None
-        pipeline = object.__new__(Cosmos3OmniDiffusersPipeline)
-        nn.Module.__init__(pipeline)
-        pipeline.transformer = transformer
-
-        source_root = tmp_path / "partition"
-        checkpoint_dir = source_root / "transformer"
-        checkpoint_dir.mkdir(parents=True)
-        checkpoint_file = checkpoint_dir / "model.safetensors"
-        save_file({"proj_in.weight": torch.ones(2, 2)}, str(checkpoint_file))
-        sources = (
-            SimpleNamespace(
-                model_or_path=str(source_root),
-                subfolder="transformer",
-                revision=None,
-                prefix="transformer.",
+        backend = DistributedLayerwiseOffloadBackend(
+            OffloadConfig(
+                strategy=OffloadStrategy.DISTRIBUTED_LAYER_WISE,
+                pin_cpu_memory=False,
+                model_path=str(tmp_path),
            ),
+            torch.device("cpu"),
        )
+        modules = SimpleNamespace(dits=[pipeline.transformer], dit_names=["transformer"])

-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", transformer),),
-            sources=sources,
-            model_path=None,
-            tensor_parallel_size=1,
-            use_hsdp=False,
-            online_quantization=False,
-        )
-
-        assert result.plan is not None
-        assert result.plan.bindings["transformer.proj_in.weight"] == TensorBinding(
-            checkpoint_key="proj_in.weight",
-            file_path=str(checkpoint_file),
-        )
+        # Should NOT raise — TP=1 is allowed even with custom weight_loader
+        backend._load_weights_via_mmap(pipeline, modules)

    def test_validate_loaded_weights_called(self, tmp_path, patched_offload_runtime):
        """validate_loaded_weights should be called after mmap loading."""
+        import json
+        from types import SimpleNamespace
+
+        from safetensors.torch import save_file
+
+        from vllm_omni.diffusion.offloader.base import OffloadConfig, OffloadStrategy

        class ModelWithValidate(nn.Module):
            _layerwise_offload_blocks_attrs = ["blocks"]
@@ -1428,64 +985,19 @@ class TestMmapValidation:

        backend = DistributedLayerwiseOffloadBackend(
            OffloadConfig(
-                strategy=OffloadStrategy.DISTRIBUTED_LAYER_WISE,
-                pin_cpu_memory=False,
+                strategy=OffloadStrategy.DISTRIBUTED_LAYER_WISE, pin_cpu_memory=False, model_path=str(tmp_path)
            ),
            torch.device("cpu"),
        )
        modules = SimpleNamespace(dits=[pipeline.transformer], dit_names=["transformer"])
-        result = build_checkpoint_mmap_plan(
-            pipeline,
-            dit_modules=(("transformer", pipeline.transformer),),
-            sources=(),
-            model_path=str(tmp_path),
-            tensor_parallel_size=1,
-            use_hsdp=False,
-            online_quantization=False,
-        )

-        assert result.plan is not None
-        backend._load_weights_via_mmap(pipeline, modules, result.plan)
+        backend._load_weights_via_mmap(pipeline, modules)
        assert pipeline.transformer.validate_called, "validate_loaded_weights should be called"


 class TestConfigValidation:
    """Tests for configuration validation in OffloadConfig / execute_request."""

-    def test_loader_plan_cannot_be_silently_dropped_on_unsupported_platform(self, monkeypatch):
-        import vllm_omni.diffusion.offloader as offloader_module
-
-        od_config = SimpleNamespace(
-            enable_cpu_offload=False,
-            enable_layerwise_offload=False,
-            enable_distributed_layerwise_offload=True,
-            dlo_use_allgather=False,
-            dlo_resident_layers=0,
-            pin_cpu_memory=False,
-            model="unused",
-            parallel_config=SimpleNamespace(
-                use_hsdp=False,
-                data_parallel_size=1,
-                sequence_parallel_size=1,
-            ),
-        )
-        plan = HostWeightPlan(
-            backing_kind="checkpoint_mmap",
-            bindings={},
-        )
-        monkeypatch.setattr(
-            offloader_module.current_omni_platform,
-            "supports_cpu_offload",
-            lambda: False,
-        )
-
-        with pytest.raises(RuntimeError, match="skipped ordinary weight materialization"):
-            offloader_module.get_offload_backend(
-                od_config,
-                device=torch.device("cpu"),
-                host_weight_plan=plan,
-            )
-
    def test_hsdp_with_allgather_rejected(self):
        """HSDP + DLO + AllGather should raise ValueError (double sharding)."""
        from vllm_omni.diffusion.offloader.base import OffloadConfig
@@ -79,11 +79,11 @@ def dist_group():


@pytest.fixture
-def patched_offload_runtime(monkeypatch):
-    monkeypatch.setattr(layerwise_backend_module.current_omni_platform, "Stream", DummyStream)
-    monkeypatch.setattr(layerwise_backend_module.current_omni_platform, "Event", DummyEvent)
-    monkeypatch.setattr(layerwise_backend_module.current_omni_platform, "current_stream", lambda: DummyStream())
-    monkeypatch.setattr(layerwise_backend_module.current_omni_platform, "stream", dummy_stream)
+def patched_offload_runtime(mocker):
+    mocker.patch.object(layerwise_backend_module.current_omni_platform, "Stream", DummyStream)
+    mocker.patch.object(layerwise_backend_module.current_omni_platform, "Event", DummyEvent)
+    mocker.patch.object(layerwise_backend_module.current_omni_platform, "current_stream", lambda: DummyStream())
+    mocker.patch.object(layerwise_backend_module.current_omni_platform, "stream", dummy_stream)


 class TinyBlock(nn.Module):
@@ -128,33 +128,6 @@ class TestLayerwiseOffloadHook:
        assert current_block.weight.to_local().shape == torch.Size([4])
        assert not hook.is_materialized

-    def test_prefetch_preserves_transposed_weight_stride(self, patched_offload_runtime):
-        """Online-FP8 Cutlass weights must retain their transposed layout."""
-
-        class StridedBlock(nn.Module):
-            def __init__(self, start: float):
-                super().__init__()
-                base = torch.arange(start, start + 12).reshape(3, 4)
-                self.weight = nn.Parameter(base.t(), requires_grad=False)
-
-        current_block = StridedBlock(1.0)
-        next_block = StridedBlock(20.0)
-        expected = next_block.weight.detach().clone()
-        expected_stride = next_block.weight.stride()
-        assert expected_stride == (1, 4)
-
-        hook = LayerwiseOffloadHook(
-            next_block=next_block,
-            device=torch.device("cpu"),
-            stream=DummyStream(),
-            pin_memory=False,
-        )
-        hook.initialize_hook(current_block)
-        hook.prefetch_layer(non_blocking=False)
-
-        assert next_block.weight.stride() == expected_stride
-        assert torch.equal(next_block.weight, expected)
-

 class _DummyBlock(nn.Module):
    def __init__(self):
@@ -16,11 +16,10 @@ from vllm_omni.model_executor.models.qwen3_omni.qwen3_omni_moe_thinker import (
 )
 from vllm_omni.quantization.component_config import (
    ComponentQuantizationConfig,
-    resolve_component_quant_config,
 )
 from vllm_omni.quantization.inc_config import OmniINCConfig

-pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
+pytestmark = [pytest.mark.core_model]


 # ---------------------------------------------------------------------------
@@ -348,24 +347,6 @@ class TestTalkerVisualRouting:
 # ===================================================================


-class TestResolveComponentQuantConfig:
-    def test_global_config_applies_to_each_quantization_aware_component(self):
-        config = _MockQuantConfig("fp8")
-
-        assert resolve_component_quant_config(config, "transformer") is config
-        assert resolve_component_quant_config(config, "text_encoder") is config
-
-    def test_component_map_is_the_only_scope_override(self):
-        transformer = _MockQuantConfig("fp8")
-        config = ComponentQuantizationConfig(
-            component_configs={"transformer": transformer},
-            default_config=None,
-        )
-
-        assert resolve_component_quant_config(config, "transformer") is transformer
-        assert resolve_component_quant_config(config, "text_encoder") is None
-
-
 class TestComponentResolve:
    def test_longest_prefix_match(self):
        a = _MockQuantConfig("a")
@@ -3,8 +3,6 @@
 """Unit tests for WorkerProc async output logic."""

 import queue
-import threading
-import time
 from unittest.mock import MagicMock

 import pytest
@@ -31,59 +29,9 @@ def _make_worker_proc(step_execution=False):
    proc._running = True
    proc._async_output_queue = queue.Queue() if not step_execution else None
    proc._async_output_thread = None
-    proc._async_output_done = threading.Condition()
-    proc._async_output_pending = 0
-    proc._result_mq_lock = threading.Lock()
    return proc


-class _OverlapDetectingQueue:
-    """Fails if two threads are inside enqueue() at the same time.
-
-    Mirrors vLLM's MessageQueue, whose acquire_write() reads current_idx,
-    writes that block, then advances it. Overlapping writers can target the
-    same ring block and silently drop a message.
-    """
-
-    def __init__(self):
-        self._active = 0
-        self._guard = threading.Lock()
-        self.overlaps = 0
-        self.messages = []
-
-    def enqueue(self, msg, timeout=None):
-        with self._guard:
-            self._active += 1
-            if self._active > 1:
-                self.overlaps += 1
-        time.sleep(0.001)  # widen the window a real ring-buffer write would have
-        with self._guard:
-            self.messages.append(msg)
-            self._active -= 1
-
-
-class TestResultQueueSingleWriter:
-    """Worker must serialize writes to the single-writer result queue."""
-
-    def test_concurrent_enqueues_do_not_overlap(self):
-        proc = _make_worker_proc(step_execution=False)
-        detector = _OverlapDetectingQueue()
-        proc.result_mq = detector
-
-        def writer(tag):
-            for i in range(20):
-                proc._enqueue_result(f"{tag}-{i}")
-
-        threads = [threading.Thread(target=writer, args=(tag,)) for tag in ("compute", "output")]
-        for t in threads:
-            t.start()
-        for t in threads:
-            t.join()
-
-        assert detector.overlaps == 0, f"{detector.overlaps} overlapping writes to result_mq"
-        assert len(detector.messages) == 40
-
-
 class TestGenerateAsyncOutputId:
    """Test _generate_async_output_id uniqueness."""

@@ -244,87 +192,6 @@ class TestAsyncOutputLoopLogic:
        assert output_ready_enqueued, "Expected OUTPUT_READY with async_output_id='abc123' in enqueue calls"


-class TestDrainAsyncOutputs:
-    """Async outputs must finish packing before device memory is released."""
-
-    def test_drain_returns_immediately_when_nothing_in_flight(self):
-        proc = _make_worker_proc(step_execution=False)
-
-        assert proc.drain_async_outputs(timeout=0.01) is True
-
-    def test_return_result_keeps_output_pending_until_loop_completes(self, mocker):
-        """_return_result marks work pending; the bg loop clears it."""
-        proc = _make_worker_proc(step_execution=False)
-
-        mock_platform = mocker.MagicMock()
-        mock_platform.record_device_event.return_value = MagicMock()
-        mocker.patch(
-            "vllm_omni.diffusion.worker.diffusion_worker.current_omni_platform",
-            mock_platform,
-        )
-        mocker.patch("vllm_omni.diffusion.worker.diffusion_worker.pack_diffusion_output_shm")
-        mock_torch = mocker.MagicMock()
-        mock_torch.accelerator.current_accelerator.return_value.type = "cpu"
-        mocker.patch("vllm_omni.diffusion.worker.diffusion_worker.torch", mock_torch)
-
-        proc._return_result(DiffusionOutput(output="data"), rpc_id="1")
-
-        # Packing has not run yet, so a sleep here would free tensors in flight.
-        assert proc._async_output_pending == 1
-        assert proc.drain_async_outputs(timeout=0.05) is False
-
-        thread = threading.Thread(target=proc._async_output_loop, daemon=True)
-        thread.start()
-        try:
-            assert proc.drain_async_outputs(timeout=5.0) is True
-        finally:
-            proc._async_output_queue.put(None)
-            thread.join(timeout=2.0)
-
-    def test_sleep_rpc_drains_before_releasing_memory(self, mocker):
-        proc = _make_worker_proc(step_execution=False)
-        proc.worker = MagicMock()
-
-        call_order: list[str] = []
-        mocker.patch.object(
-            proc,
-            "drain_async_outputs",
-            side_effect=lambda *a, **kw: call_order.append("drain") or True,
-        )
-        proc.worker.execute_method.side_effect = lambda *a, **kw: call_order.append("sleep")
-
-        proc._execute_rpc(
-            {
-                "method": "handle_sleep_task",
-                "args": (),
-                "kwargs": {},
-                "output_rank": 0,
-                "exec_all_ranks": False,
-                "collect_rank_status": False,
-            }
-        )
-
-        assert call_order == ["drain", "sleep"]
-
-    def test_non_sleep_rpc_does_not_drain(self, mocker):
-        proc = _make_worker_proc(step_execution=False)
-        proc.worker = MagicMock()
-        drain = mocker.patch.object(proc, "drain_async_outputs", return_value=True)
-
-        proc._execute_rpc(
-            {
-                "method": "execute_model",
-                "args": (),
-                "kwargs": {},
-                "output_rank": 0,
-                "exec_all_ranks": False,
-                "collect_rank_status": False,
-            }
-        )
-
-        drain.assert_not_called()
-
-
 class TestWorkerProcShutdown:
    def test_shutdown_stops_async_thread_and_releases_queues(self):
        import threading
@@ -123,7 +123,6 @@ class _ConcurrencyTrackingExecutor:
 def _make_request(tag: str):
    sampling_params = dict(_SAMPLING_KEY_DEFAULTS)
    sampling_params["num_inference_steps"] = 1
-    sampling_params["extra_args"] = {}
    return SimpleNamespace(
        request_id=tag,
        prompt=f"prompt_{tag}",
@@ -305,49 +305,6 @@ def test_batch_runner_output_round_trips_nested_results_through_shm() -> None:
    assert output["req-error"].result.error == "boom"


-def test_per_request_views_of_a_batch_tensor_do_not_ship_whole_storage() -> None:
-    """Regression: per-request slices used to pickle the whole batch storage.
-
-    Each view is under the packing threshold, but pickle writes the entire
-    shared storage per view, so a batch's message grew by the full batch
-    tensor once per request and overflowed the ring buffer.
-    """
-    import pickle
-
-    num_requests = 16
-    # Each row stays under the threshold; the shared storage is well over it.
-    row_numel = _SHM_TENSOR_THRESHOLD // (2 * torch.empty((), dtype=torch.float32).element_size())
-    batch = torch.arange(num_requests * row_numel, dtype=torch.float32).reshape(num_requests, row_numel)
-    per_request = batch[0]
-    assert per_request.nelement() * per_request.element_size() <= _SHM_TENSOR_THRESHOLD
-    assert per_request.untyped_storage().nbytes() > _SHM_TENSOR_THRESHOLD
-
-    output = BatchRunnerOutput.from_list(
-        [
-            RunnerOutput(
-                request_id=f"req-{i}",
-                finished=True,
-                result=DiffusionOutput(trajectory_latents=batch[i]),
-            )
-            for i in range(num_requests)
-        ]
-    )
-    unpacked_size = len(pickle.dumps(output, protocol=pickle.HIGHEST_PROTOCOL))
-
-    pack_diffusion_output_shm(output)
-    packed_size = len(pickle.dumps(output, protocol=pickle.HIGHEST_PROTOCOL))
-
-    assert output.runner_outputs[0].result.trajectory_latents["__tensor_shm__"] is True
-    # Whole storage per request before; handles only after.
-    storage_bytes = per_request.untyped_storage().nbytes()
-    assert unpacked_size > num_requests * storage_bytes / 2
-    assert packed_size < _SHM_TENSOR_THRESHOLD
-
-    unpack_diffusion_output_shm(output)
-    for i in range(num_requests):
-        torch.testing.assert_close(output[f"req-{i}"].result.trajectory_latents, batch[i])
-
-
 def test_pack_value_keeps_tensor_at_threshold_inline() -> None:
    tensor = torch.arange(
        _SHM_TENSOR_THRESHOLD // torch.empty((), dtype=torch.float32).element_size(),
@@ -740,9 +740,6 @@ def test_load_model_clears_cache_backend_for_unsupported_pipeline(monkeypatch):
            del kwargs
            return SimpleNamespace(transformer=torch.nn.Identity())

-        def take_host_weight_plan(self):
-            return None
-
    class _DummyMemoryProfiler:
        consumed_memory = 0

@@ -782,11 +779,7 @@ def test_load_model_clears_cache_backend_for_unsupported_pipeline(monkeypatch):
    monkeypatch.setattr(model_runner_module, "LoadConfig", lambda: object())
    monkeypatch.setattr(model_runner_module, "DiffusersPipelineLoader", _DummyLoader)
    monkeypatch.setattr(model_runner_module, "DeviceMemoryProfiler", _DummyMemoryProfiler)
-    monkeypatch.setattr(
-        model_runner_module,
-        "get_offload_backend",
-        lambda od_config, device, host_weight_plan: None,
-    )
+    monkeypatch.setattr(model_runner_module, "get_offload_backend", lambda od_config, device: None)
    monkeypatch.setattr(
        model_runner_module, "get_cache_backend", lambda cache_backend, cache_config: dummy_cache_backend
    )
@@ -217,21 +217,13 @@ class TestGetRequestBatchSamplingParamsKey:
        num_inference_steps: int = 2,
        seed: int | None = 123,
        generator: torch.Generator | None = None,
-        extra_args: dict | None = None,
-        condition_key: tuple | None = None,
    ) -> OmniDiffusionRequest:
        sp = OmniDiffusionSamplingParams(
            num_inference_steps=num_inference_steps,
            seed=seed,
            generator=generator,
-            extra_args=extra_args or {},
-        )
-        return OmniDiffusionRequest(
-            prompt="prompt",
-            sampling_params=sp,
-            request_id=f"req-{num_inference_steps}",
-            batch_compatibility_key=condition_key,
        )
+        return OmniDiffusionRequest(prompt="prompt", sampling_params=sp, request_id=f"req-{num_inference_steps}")

    def test_distinguishes_num_inference_steps(self) -> None:
        scheduler = RequestScheduler()
@@ -248,57 +240,6 @@ class TestGetRequestBatchSamplingParamsKey:
            self._make(seed=1, generator=gen_a)
        ) == scheduler._build_sampling_params_key(self._make(seed=2, generator=gen_b))

-    @pytest.mark.parametrize(
-        ("first_extra_args", "second_extra_args"),
-        [
-            ({"sample_solver": "unipc"}, {"sample_solver": "euler"}),
-            ({"flow_shift": 3.0}, {"flow_shift": 5.0}),
-        ],
-    )
-    def test_distinguishes_wan_scheduler_structure(
-        self,
-        first_extra_args: dict,
-        second_extra_args: dict,
-    ) -> None:
-        scheduler = RequestScheduler()
-
-        assert scheduler._build_sampling_params_key(
-            self._make(extra_args=first_extra_args)
-        ) != scheduler._build_sampling_params_key(self._make(extra_args=second_extra_args))
-
-    @pytest.mark.parametrize(
-        ("first_extra_args", "second_extra_args"),
-        [
-            ({"sample_solver": " Euler "}, {"sample_solver": "euler"}),
-            ({"flow_shift": "5.0"}, {"flow_shift": 5.0}),
-        ],
-    )
-    def test_normalizes_equivalent_wan_scheduler_structure(
-        self,
-        first_extra_args: dict,
-        second_extra_args: dict,
-    ) -> None:
-        scheduler = RequestScheduler()
-
-        assert scheduler._build_sampling_params_key(
-            self._make(extra_args=first_extra_args)
-        ) == scheduler._build_sampling_params_key(self._make(extra_args=second_extra_args))
-
-    def test_uses_none_for_unspecified_wan_scheduler_structure(self) -> None:
-        scheduler = RequestScheduler()
-
-        key = scheduler._build_sampling_params_key(self._make())
-
-        assert key.sample_solver is None
-        assert key.flow_shift is None
-
-    def test_distinguishes_pipeline_condition_structure(self) -> None:
-        scheduler = RequestScheduler()
-
-        assert scheduler._build_sampling_params_key(
-            self._make(condition_key=("wan22_s2v_condition", True))
-        ) != scheduler._build_sampling_params_key(self._make(condition_key=("wan22_s2v_condition", False)))
-

 class TestRequestScheduler:
    def setup_method(self) -> None:
@@ -661,38 +602,6 @@ class TestRequestScheduler:
        assert first.num_running_reqs == 1
        assert first.num_waiting_reqs == 1

-    def test_batches_incompatible_pipeline_conditions_separately(self) -> None:
-        scheduler = RequestScheduler()
-        scheduler.initialize(SimpleNamespace(max_num_seqs=2))
-
-        request_a = OmniDiffusionRequest(
-            prompt="a",
-            sampling_params=OmniDiffusionSamplingParams(num_inference_steps=2, seed=123),
-            request_id="a",
-            batch_compatibility_key=("wan22_s2v_condition", True),
-        )
-        request_b = OmniDiffusionRequest(
-            prompt="b",
-            sampling_params=OmniDiffusionSamplingParams(num_inference_steps=2, seed=456),
-            request_id="b",
-            batch_compatibility_key=("wan22_s2v_condition", False),
-        )
-        scheduler.add_request(request_a)
-        scheduler.add_request(request_b)
-
-        first = scheduler.schedule()
-
-        assert _new_ids(first) == ["a"]
-        assert first.num_running_reqs == 1
-        assert first.num_waiting_reqs == 1
-
-        scheduler.update_from_output(first, _make_request_output("a"))
-        second = scheduler.schedule()
-
-        assert _new_ids(second) == ["b"]
-        assert second.num_running_reqs == 1
-        assert second.num_waiting_reqs == 0
-
    def test_batches_different_request_local_seed_together(self) -> None:
        scheduler = RequestScheduler()
        scheduler.initialize(SimpleNamespace(max_num_seqs=2))
@@ -716,47 +625,6 @@ class TestRequestScheduler:
        assert first.num_running_reqs == 2
        assert first.num_waiting_reqs == 0

-    @pytest.mark.parametrize(
-        ("first_extra_args", "second_extra_args"),
-        [
-            ({"sample_solver": "unipc"}, {"sample_solver": "euler"}),
-            ({"flow_shift": 3.0}, {"flow_shift": 5.0}),
-            ({"sample_solver": "unipc"}, {}),
-            ({"flow_shift": 3.0}, {}),
-        ],
-    )
-    def test_batches_incompatible_wan_scheduler_structure_separately(
-        self,
-        first_extra_args: dict,
-        second_extra_args: dict,
-    ) -> None:
-        scheduler = RequestScheduler()
-        scheduler.initialize(SimpleNamespace(max_num_seqs=2))
-        first_id = scheduler.add_request(
-            _make_step_request(
-                "a",
-                sampling_params=OmniDiffusionSamplingParams(
-                    num_inference_steps=2,
-                    extra_args=first_extra_args,
-                ),
-            )
-        )
-        scheduler.add_request(
-            _make_step_request(
-                "b",
-                sampling_params=OmniDiffusionSamplingParams(
-                    num_inference_steps=2,
-                    extra_args=second_extra_args,
-                ),
-            )
-        )
-
-        first = scheduler.schedule()
-
-        assert _new_ids(first) == [first_id]
-        assert first.num_running_reqs == 1
-        assert first.num_waiting_reqs == 1
-
    def test_incompatible_waiting_head_blocks_later_compatible_request(self) -> None:
        scheduler = RequestScheduler()
        scheduler.initialize(SimpleNamespace(max_num_seqs=3))
@@ -1,171 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-
-import threading
-import time
-from unittest.mock import Mock
-
-import pytest
-from vllm.distributed.device_communicators.shm_broadcast import MessageQueue
-
-import vllm_omni.diffusion.worker.diffusion_worker as diffusion_worker_module
-from vllm_omni.diffusion.worker.diffusion_worker import WorkerProc
-
-pytestmark = [pytest.mark.core_model, pytest.mark.diffusion, pytest.mark.cpu]
-
-
-class _DelayedMultipartSocket:
-    def __init__(
-        self,
-        socket,
-        send_started: threading.Event,
-        release_send: threading.Event,
-    ):
-        self._socket = socket
-        self._send_started = send_started
-        self._release_send = release_send
-        self._delay_next_send = True
-
-    def __getattr__(self, name):
-        return getattr(self._socket, name)
-
-    def send_multipart(self, *args, **kwargs):
-        if self._delay_next_send:
-            self._delay_next_send = False
-            self._send_started.set()
-            if not self._release_send.wait(timeout=2):
-                raise TimeoutError("Timed out waiting to release delayed overflow payload")
-        return self._socket.send_multipart(*args, **kwargs)
-
-
-class _ObservedPollSocket:
-    def __init__(self, socket, poll_started: threading.Event):
-        self._socket = socket
-        self._poll_started = poll_started
-
-    def __getattr__(self, name):
-        return getattr(self._socket, name)
-
-    def poll(self, *args, **kwargs):
-        self._poll_started.set()
-        return self._socket.poll(*args, **kwargs)
-
-
-class _RecordingWorker:
-    def __init__(self):
-        self.calls: list[str] = []
-
-    def execute_method(self, method, *args, **kwargs):
-        assert method == "record"
-        self.calls.append(args[0])
-
-
-def test_recv_message_waits_indefinitely():
-    worker_proc = WorkerProc.__new__(WorkerProc)
-    worker_proc.mq = Mock()
-    worker_proc.mq.dequeue.return_value = {"type": "rpc"}
-
-    result = worker_proc.recv_message()
-
-    assert result == {"type": "rpc"}
-    worker_proc.mq.dequeue.assert_called_once_with(indefinite=True)
-
-
-def test_busy_loop_preserves_order_when_overflow_payload_is_delayed(monkeypatch):
-    """A consumed overflow marker must not be abandoned on a socket timeout."""
-    writer = MessageQueue(
-        n_reader=1,
-        n_local_reader=1,
-        local_reader_ranks=[0],
-        max_chunk_bytes=1024,
-        max_chunks=4,
-    )
-    reader = MessageQueue.create_from_handle(writer.export_handle(), rank=0)
-    writer.wait_until_ready()
-    reader.wait_until_ready()
-
-    overflow_send_started = threading.Event()
-    release_overflow_send = threading.Event()
-    overflow_poll_started = threading.Event()
-    writer.local_socket = _DelayedMultipartSocket(
-        writer.local_socket,
-        overflow_send_started,
-        release_overflow_send,
-    )
-    reader.local_socket = _ObservedPollSocket(reader.local_socket, overflow_poll_started)
-
-    recording_worker = _RecordingWorker()
-    worker_proc = WorkerProc.__new__(WorkerProc)
-    worker_proc.gpu_id = 0
-    worker_proc.mq = reader
-    worker_proc.worker = recording_worker
-    worker_proc.result_mq = None
-    worker_proc.wake_event = None
-    worker_proc._running = True
-
-    # Make the pre-fix implementation fail quickly: it read this module-level
-    # value directly from _worker_busy_loop. The fixed implementation ignores it
-    # and delegates to recv_message(), which waits indefinitely.
-    monkeypatch.setattr(
-        diffusion_worker_module,
-        "_BROADCAST_DEQUEUE_TIMEOUT_S",
-        0.02,
-        raising=False,
-    )
-
-    overflow_message = {
-        "type": "rpc",
-        "method": "record",
-        "args": ("overflow-1",),
-        "padding": b"x" * 2048,
-    }
-    inline_message = {
-        "type": "rpc",
-        "method": "record",
-        "args": ("inline-2",),
-    }
-    shutdown_message = {"type": "shutdown"}
-    thread_errors: list[BaseException] = []
-
-    def send_messages():
-        try:
-            writer.enqueue(overflow_message)
-            writer.enqueue(inline_message)
-            writer.enqueue(shutdown_message)
-        except BaseException as exc:
-            thread_errors.append(exc)
-
-    def run_busy_loop():
-        try:
-            worker_proc._worker_busy_loop()
-        except BaseException as exc:
-            thread_errors.append(exc)
-
-    writer_thread = threading.Thread(target=send_messages, daemon=True)
-    worker_thread = threading.Thread(target=run_busy_loop, daemon=True)
-
-    try:
-        writer_thread.start()
-        worker_thread.start()
-
-        assert overflow_send_started.wait(timeout=2)
-        assert overflow_poll_started.wait(timeout=2)
-
-        # Keep the multipart unavailable beyond the pre-fix 20 ms timeout.
-        time.sleep(0.1)
-        release_overflow_send.set()
-
-        writer_thread.join(timeout=2)
-        worker_thread.join(timeout=2)
-
-        assert not writer_thread.is_alive()
-        assert not worker_thread.is_alive()
-        assert thread_errors == []
-        assert recording_worker.calls == ["overflow-1", "inline-2"]
-    finally:
-        release_overflow_send.set()
-        worker_proc._running = False
-        reader.shutdown()
-        writer.shutdown()
-        writer_thread.join(timeout=1)
-        worker_thread.join(timeout=1)
@@ -6,7 +6,6 @@ import concurrent.futures
 import queue
 import threading
 import time
-from types import SimpleNamespace
 from unittest.mock import MagicMock

 import pytest
@@ -265,119 +264,6 @@ class TestResultPumpDispatch:
        fut = executor.wait_output_ready(async_output_id)
        assert fut.done()
        assert fut.result(timeout=1.0) is output
-        assert async_output_id not in executor._output_futures
-
-    def test_output_ready_atomic_resolution_when_future_already_waiting(self):
-        """When OUTPUT_READY arrives and a future is already waiting, resolve it directly."""
-        executor = _make_executor()
-        async_output_id = "abc123"
-        output = DiffusionOutput(output="waiting")
-        fut = executor.wait_output_ready(async_output_id)
-        assert not fut.done()
-
-        msg = AsyncDiffusionOutput(
-            kind=AsyncOutputKind.OUTPUT_READY,
-            async_output_id=async_output_id,
-            output=output,
-        )
-        _feed_one_msg_to_pump(executor, msg)
-
-        assert fut.done()
-        assert fut.result(timeout=1.0) is output
-        assert async_output_id not in executor._output_futures
-        assert async_output_id not in executor._completed_outputs
-
-
-class _FakeBatchOutput:
-    """Batch-level output exposing per-request results."""
-
-    def __init__(self, results):
-        self._results = results
-
-    def get_request_output(self, req_id):
-        result = self._results.get(req_id)
-        if result is None:
-            return None
-        return SimpleNamespace(result=result)
-
-
-def _make_scheduler_output(req_ids):
-    return SimpleNamespace(
-        scheduled_new_reqs=[SimpleNamespace(request_id=rid, req=SimpleNamespace()) for rid in req_ids]
-    )
-
-
-class TestBatchSplitDelivery:
-    """execute_batch must resolve per-request futures in either arrival order."""
-
-    @staticmethod
-    def _run(executor, req_ids, batch_id, deliver_early):
-        # Keep execute_batch on the fused request-batch path (not DLO DP).
-        executor.od_config.parallel_config.data_parallel_size = 1
-        executor.od_config.enable_distributed_layerwise_offload = False
-        executor._ensure_open = lambda: None
-
-        outputs = {rid: DiffusionOutput(output=f"img-{rid}") for rid in req_ids}
-        ready = AsyncDiffusionOutput(
-            kind=AsyncOutputKind.OUTPUT_READY,
-            async_output_id=batch_id,
-            output=_FakeBatchOutput(outputs),
-        )
-
-        def fake_collective_rpc(*args, **kwargs):
-            if deliver_early:
-                # Worker's background D2H/SHM thread wins the race: OUTPUT_READY
-                # is pumped before execute_batch registers the split map.
-                _feed_one_msg_to_pump(executor, ready)
-            return AsyncDiffusionOutput(
-                kind=AsyncOutputKind.COMPUTE_DONE,
-                rpc_id="1",
-                async_output_id=batch_id,
-            )
-
-        executor.collective_rpc = fake_collective_rpc
-        batch = executor.execute_batch(_make_scheduler_output(req_ids))
-        if not deliver_early:
-            _feed_one_msg_to_pump(executor, ready)
-        return batch, outputs
-
-    def test_output_ready_after_split_map(self):
-        executor = _make_executor()
-        req_ids = ["r0", "r1", "r2"]
-        _, outputs = self._run(executor, req_ids, "batch-1", deliver_early=False)
-
-        for rid in req_ids:
-            fut = executor.wait_output_ready(f"batch-1/{rid}")
-            assert fut.done()
-            assert fut.result(timeout=1.0) is outputs[rid]
-
-    def test_output_ready_before_split_map(self):
-        """Regression: the whole batch used to be lost, hanging every request."""
-        executor = _make_executor()
-        req_ids = ["r0", "r1", "r2"]
-        _, outputs = self._run(executor, req_ids, "batch-1", deliver_early=True)
-
-        for rid in req_ids:
-            fut = executor.wait_output_ready(f"batch-1/{rid}")
-            assert fut.done(), f"request {rid} never resolved"
-            assert fut.result(timeout=1.0) is outputs[rid]
-
-        # No stale batch-level state left behind.
-        assert executor._batch_split_map == {}
-        assert executor._completed_outputs == {}
-
-    def test_engine_waiter_before_output_is_resolved(self):
-        """Consumers already blocked in wait_output_ready must be woken."""
-        executor = _make_executor()
-        req_ids = ["r0", "r1"]
-        batch_id = "batch-2"
-        futures = {rid: executor.wait_output_ready(f"{batch_id}/{rid}") for rid in req_ids}
-
-        _, outputs = self._run(executor, req_ids, batch_id, deliver_early=True)
-
-        for rid in req_ids:
-            assert futures[rid].done(), f"request {rid} never resolved"
-            assert futures[rid].result(timeout=1.0) is outputs[rid]


 class TestShutdownCleansUpFutures:
@@ -1,55 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""Guard the public accessor that examples and docs use to read omni outputs.
-
-``OmniRequestOutput`` used to hold a nested ``RequestOutput`` in a
-``request_output`` field.  Since #5146 it *inherits* ``RequestOutput``, so the
-generation content lives on the object itself and the nested field is gone.
-
-Call sites that kept the old accessor only fail once a real model runs, which
-means a GPU nightly job is the first thing to notice.  These checks pin both
-halves of the contract on CPU instead: the field is really gone, and no doc,
-example, test or tool still reaches for it.
-"""
-
-import re
-from pathlib import Path
-
-import pytest
-
-from vllm_omni.outputs import OmniRequestOutput
-
-pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
-
-
-ROOT_DIR = Path(__file__).parents[2]
-SCANNED_DIRS = ("docs", "examples", "tests", "tools")
-SCANNED_SUFFIXES = (".py", ".md")
-
-# ``.request_outputs`` (plural, an OutputProcessor batch) is a different and
-# still-valid attribute, so require a non-identifier character after the name.
-NESTED_ACCESSOR = re.compile(r"\.request_output(?![\w])")
-
-
-def _iter_scanned_files():
-    for directory in SCANNED_DIRS:
-        for path in sorted((ROOT_DIR / directory).rglob("*")):
-            if path.suffix in SCANNED_SUFFIXES and path.resolve() != Path(__file__).resolve():
-                yield path
-
-
-def test_omni_request_output_has_no_nested_request_output():
-    assert not hasattr(OmniRequestOutput(request_id="req-accessor"), "request_output")
-
-
-def test_no_source_reads_the_nested_request_output():
-    offenders = []
-    for path in _iter_scanned_files():
-        for lineno, line in enumerate(path.read_text(encoding="utf-8").splitlines(), start=1):
-            if NESTED_ACCESSOR.search(line):
-                offenders.append(f"{path.relative_to(ROOT_DIR)}:{lineno}: {line.strip()}")
-
-    assert not offenders, (
-        "OmniRequestOutput no longer nests a RequestOutput; read the content off the "
-        "output itself (e.g. `outputs[0].images`) instead of `.request_output`:\n" + "\n".join(offenders)
-    )
@@ -1,591 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-
-"""Generate raw LTX video and audio outputs with the official or Omni runtime."""
-
-from __future__ import annotations
-
-import argparse
-import json
-import os
-import sys
-from pathlib import Path
-from typing import Any
-
-import numpy as np
-import torch
-from PIL import Image
-
-
-def _parse_args() -> argparse.Namespace:
-    parser = argparse.ArgumentParser()
-    parser.add_argument("--backend", choices=("official", "omni"), required=True)
-    parser.add_argument("--request", type=Path, required=True)
-    parser.add_argument("--output-dir", type=Path, required=True)
-    parser.add_argument("--model")
-    parser.add_argument("--model-class-name")
-    parser.add_argument("--official-root", type=Path)
-    parser.add_argument("--checkpoint", type=Path)
-    parser.add_argument("--gemma-root", type=Path)
-    parser.add_argument("--transformer-path", type=Path)
-    parser.add_argument("--text-encoder-path", type=Path)
-    parser.add_argument("--video-vae-path", type=Path)
-    parser.add_argument("--audio-vae-path", type=Path)
-    parser.add_argument("--spatial-upsampler-path", type=Path)
-    parser.add_argument("--distilled-lora-path", type=Path)
-    parser.add_argument("--connector-model", type=Path)
-    parser.add_argument(
-        "--official-pipeline",
-        choices=("distilled_two_stage", "full_one_stage", "full_two_stage"),
-        default="distilled_two_stage",
-    )
-    parser.add_argument("--enable-layerwise-offload", action="store_true")
-    return parser.parse_args()
-
-
-def _save_outputs(
-    output_dir: Path,
-    *,
-    video: torch.Tensor,
-    audio: torch.Tensor,
-    audio_sample_rate: int,
-    fps: float,
-    metadata: dict[str, Any],
-) -> None:
-    output_dir.mkdir(parents=True, exist_ok=True)
-    video = video.detach().float().cpu().clamp(0.0, 1.0)
-    audio = audio.detach().float().cpu()
-    np.save(output_dir / "video.npy", video.numpy())
-    np.save(output_dir / "audio.npy", audio.numpy())
-
-    from vllm_omni.diffusion.utils.media_utils import mux_video_audio_bytes
-
-    frames_u8 = video.mul(255.0).round().to(torch.uint8).numpy()
-    audio_array = audio.numpy()
-    while audio_array.ndim > 2 and audio_array.shape[0] == 1:
-        audio_array = audio_array[0]
-    (output_dir / "output.mp4").write_bytes(
-        mux_video_audio_bytes(
-            frames_u8,
-            audio_array,
-            fps=float(fps),
-            audio_sample_rate=int(audio_sample_rate),
-            video_codec_options={"preset": "ultrafast"},
-        )
-    )
-
-    frame_indices = sorted({0, video.shape[0] // 2, video.shape[0] - 1})
-    for index in frame_indices:
-        frame = video[index].mul(255.0).round().to(torch.uint8).numpy()
-        Image.fromarray(frame).save(output_dir / f"frame_{index:04d}.png")
-
-    metadata.update(
-        {
-            "video_shape": list(video.shape),
-            "audio_shape": list(audio.shape),
-            "audio_sample_rate": int(audio_sample_rate),
-            "frame_indices": frame_indices,
-            "mp4": "output.mp4",
-        }
-    )
-    (output_dir / "metadata.json").write_text(json.dumps(metadata, indent=2) + "\n")
-
-
-def _reset_accelerator_peak_memory_stats() -> None:
-    if torch.accelerator.is_available():
-        torch.accelerator.reset_peak_memory_stats()
-
-
-def _accelerator_peak_memory_stats() -> dict[str, float]:
-    if not torch.accelerator.is_available():
-        return {}
-    torch.accelerator.synchronize()
-    mib = 1024**2
-    return {
-        "peak_memory_allocated_mb": torch.accelerator.max_memory_allocated() / mib,
-        "peak_memory_reserved_mb": torch.accelerator.max_memory_reserved() / mib,
-    }
-
-
-def _insert_official_paths(official_root: Path) -> None:
-    for relative_path in ("packages/ltx-core/src", "packages/ltx-pipelines/src"):
-        path = str((official_root / relative_path).resolve())
-        if path not in sys.path:
-            sys.path.insert(0, path)
-
-
-def _configure_official_sdpa(pipeline: Any) -> None:
-    """Pin official connector and denoiser attention to cuDNN, with MATH fallback."""
-    from ltx_core.loader.attention_ops import set_attention_module_op
-    from ltx_core.model.transformer.attention import PytorchAttention
-    from torch.nn.attention import SDPBackend
-
-    attention = PytorchAttention(priority=[SDPBackend.CUDNN_ATTENTION, SDPBackend.MATH])
-    module_op = set_attention_module_op(
-        attention=attention,
-        masked_attention=attention,
-    )
-    transformer_owners = [
-        owner
-        for attribute in ("stage", "stage_1", "stage_2")
-        if (owner := getattr(pipeline, attribute, None)) is not None
-    ]
-    owners_and_attributes = [
-        *((owner, "_transformer_builder") for owner in transformer_owners),
-        (pipeline.prompt_encoder, "_embeddings_processor_builder"),
-    ]
-    for owner, attribute in owners_and_attributes:
-        builder = getattr(owner, attribute)
-        setattr(
-            owner,
-            attribute,
-            builder.with_module_ops((*builder.module_ops, module_op)),
-        )
-
-
-def _use_diffusers_connector_weights(prompt_encoder: Any, model: Path) -> None:
-    """Run official connector code with the checkpoint under test."""
-    from safetensors import safe_open
-
-    connector_root = model / "connectors"
-    shards = sorted(connector_root.glob("*.safetensors"))
-    if not shards:
-        raise ValueError(f"No connector safetensors found under {connector_root}")
-
-    original_build_embeddings_processor = prompt_encoder._build_embeddings_processor
-
-    def build_embeddings_processor():
-        processor = original_build_embeddings_processor()
-        parameters = dict(processor.named_parameters())
-        expected = {name for name in parameters if name.startswith(("video_connector.", "audio_connector."))}
-        loaded: set[str] = set()
-        with torch.no_grad():
-            for shard in shards:
-                with safe_open(str(shard), framework="pt", device="cpu") as weights:
-                    for source_name in weights.keys():
-                        if not source_name.startswith(("video_connector.", "audio_connector.")):
-                            continue
-                        target_name = (
-                            source_name.replace(".transformer_blocks.", ".transformer_1d_blocks.")
-                            .replace(".attn1.norm_q.", ".attn1.q_norm.")
-                            .replace(".attn1.norm_k.", ".attn1.k_norm.")
-                        )
-                        parameter = parameters[target_name]
-                        parameter.copy_(weights.get_tensor(source_name).to(parameter.device, parameter.dtype))
-                        loaded.add(target_name)
-        if loaded != expected:
-            missing = sorted(expected - loaded)
-            unexpected = sorted(loaded - expected)
-            raise ValueError(f"Connector weight mismatch: missing={missing}, unexpected={unexpected}")
-        return processor
-
-    prompt_encoder._build_embeddings_processor = build_embeddings_processor
-
-
-@torch.inference_mode()
-def _run_official(args: argparse.Namespace, request: dict[str, Any]) -> None:
-    if args.official_root is None:
-        raise ValueError("Official backend requires --official-root")
-    _insert_official_paths(args.official_root)
-    # Keep Gemma on the same SDPA path in both subprocesses. DiT attention is
-    # still pinned independently to the explicit cuDNN backend below.
-    torch.backends.cuda.enable_cudnn_sdp(False)
-    _reset_accelerator_peak_memory_stats()
-
-    from ltx_pipelines.utils.args import ImageConditioningInput
-    from ltx_pipelines.utils.model_paths import ModelPaths
-    from ltx_pipelines.utils.types import OffloadMode
-
-    offload_mode = OffloadMode.CPU if args.enable_layerwise_offload else OffloadMode.NONE
-    image_path = request.get("image")
-    images = (
-        []
-        if image_path is None
-        else [
-            ImageConditioningInput(
-                path=str(image_path),
-                frame_idx=0,
-                strength=1.0,
-                crf=request.get("image_crf", 18),
-            )
-        ]
-    )
-
-    if args.transformer_path is not None:
-        required = {
-            "--text-encoder-path": args.text_encoder_path,
-            "--video-vae-path": args.video_vae_path,
-            "--audio-vae-path": args.audio_vae_path,
-        }
-        if args.official_pipeline in {"distilled_two_stage", "full_two_stage"}:
-            required["--spatial-upsampler-path"] = args.spatial_upsampler_path
-        if args.official_pipeline == "full_two_stage":
-            required["--distilled-lora-path"] = args.distilled_lora_path
-        missing = [name for name, value in required.items() if value is None]
-        if missing:
-            raise ValueError(f"Official LTX-2.5 split backend is missing: {', '.join(missing)}")
-
-        model_paths = ModelPaths.from_split(
-            transformer_path=str(args.transformer_path),
-            text_encoder_path=str(args.text_encoder_path),
-            video_vae_path=str(args.video_vae_path),
-            audio_vae_path=str(args.audio_vae_path),
-        )
-        if args.official_pipeline == "distilled_two_stage":
-            from ltx_pipelines.distilled import DistilledPipeline
-
-            pipeline = DistilledPipeline(
-                model_paths=model_paths,
-                spatial_upsampler_path=str(args.spatial_upsampler_path),
-                loras=(),
-                offload_mode=offload_mode,
-            )
-            if args.connector_model is not None:
-                _use_diffusers_connector_weights(pipeline.prompt_encoder, args.connector_model)
-            _configure_official_sdpa(pipeline)
-            video, audio, _, _ = pipeline(
-                prompt=request["prompt"],
-                seed=request["seed"],
-                height=request["height"],
-                width=request["width"],
-                num_frames=request["num_frames"],
-                frame_rate=request["fps"],
-                images=images,
-            )
-            pipeline_name = "DistilledPipeline"
-        elif args.official_pipeline == "full_two_stage":
-            from ltx_core.components.guiders import MultiModalGuiderParams
-            from ltx_core.loader import LTXV_LORA_COMFY_RENAMING_MAP, LoraPathStrengthAndSDOps
-            from ltx_pipelines.ti2vid_two_stages import TI2VidTwoStagesPipeline
-
-            pipeline = TI2VidTwoStagesPipeline(
-                model_paths=model_paths,
-                distilled_lora=[
-                    LoraPathStrengthAndSDOps(
-                        path=str(args.distilled_lora_path),
-                        strength=1.0,
-                        sd_ops=LTXV_LORA_COMFY_RENAMING_MAP,
-                    )
-                ],
-                spatial_upsampler_path=str(args.spatial_upsampler_path),
-                loras=[],
-                offload_mode=offload_mode,
-            )
-            if args.connector_model is not None:
-                _use_diffusers_connector_weights(pipeline.prompt_encoder, args.connector_model)
-            _configure_official_sdpa(pipeline)
-            video, audio, _, _ = pipeline(
-                prompt=request["prompt"],
-                negative_prompt=request["negative_prompt"],
-                seed=request["seed"],
-                height=request["height"],
-                width=request["width"],
-                num_frames=request["num_frames"],
-                frame_rate=request["fps"],
-                num_inference_steps=request["num_inference_steps"],
-                video_guider_params=MultiModalGuiderParams(
-                    cfg_scale=request["video_cfg_scale"],
-                    stg_scale=request["video_stg_scale"],
-                    rescale_scale=request["video_rescale_scale"],
-                    modality_scale=request["video_modality_scale"],
-                    skip_step=0,
-                    stg_blocks=request["video_stg_blocks"],
-                ),
-                audio_guider_params=MultiModalGuiderParams(
-                    cfg_scale=request["audio_cfg_scale"],
-                    stg_scale=request["audio_stg_scale"],
-                    rescale_scale=request["audio_rescale_scale"],
-                    modality_scale=request["audio_modality_scale"],
-                    skip_step=0,
-                    stg_blocks=request["audio_stg_blocks"],
-                ),
-                images=images,
-                max_batch_size=4,
-                stage_1_sigmas=torch.tensor(request["stage_1_sigmas"], dtype=torch.float32),
-                stage_2_sigmas=torch.tensor(request["stage_2_sigmas"], dtype=torch.float32),
-            )
-            pipeline_name = "TI2VidTwoStagesPipeline"
-        else:
-            from ltx_core.components.guiders import MultiModalGuiderParams
-            from ltx_pipelines.ti2vid_one_stage import TI2VidOneStagePipeline
-
-            pipeline = TI2VidOneStagePipeline(
-                model_paths=model_paths,
-                loras=(),
-                offload_mode=offload_mode,
-            )
-            if args.connector_model is not None:
-                _use_diffusers_connector_weights(pipeline.prompt_encoder, args.connector_model)
-            _configure_official_sdpa(pipeline)
-            request_sigmas = request.get("sigmas")
-            video, audio, _ = pipeline(
-                prompt=request["prompt"],
-                negative_prompt=request["negative_prompt"],
-                seed=request["seed"],
-                height=request["height"],
-                width=request["width"],
-                num_frames=request["num_frames"],
-                frame_rate=request["fps"],
-                num_inference_steps=request["num_inference_steps"],
-                video_guider_params=MultiModalGuiderParams(
-                    cfg_scale=request["video_cfg_scale"],
-                    stg_scale=request["video_stg_scale"],
-                    rescale_scale=request["video_rescale_scale"],
-                    modality_scale=request["video_modality_scale"],
-                    skip_step=0,
-                    stg_blocks=request["video_stg_blocks"],
-                ),
-                audio_guider_params=MultiModalGuiderParams(
-                    cfg_scale=request["audio_cfg_scale"],
-                    stg_scale=request["audio_stg_scale"],
-                    rescale_scale=request["audio_rescale_scale"],
-                    modality_scale=request["audio_modality_scale"],
-                    skip_step=0,
-                    stg_blocks=request["audio_stg_blocks"],
-                ),
-                images=images,
-                max_batch_size=4,
-                sigmas=None if request_sigmas is None else torch.tensor(request_sigmas, dtype=torch.float32),
-            )
-            pipeline_name = "TI2VidOneStagePipeline"
-        checkpoint_metadata = {
-            "transformer_path": str(args.transformer_path),
-            "video_vae_path": str(args.video_vae_path),
-            "connector_model": None if args.connector_model is None else str(args.connector_model),
-            "pipeline": pipeline_name,
-        }
-    else:
-        if args.checkpoint is None or args.gemma_root is None:
-            raise ValueError("Official monolith backend requires --checkpoint and --gemma-root")
-
-        from ltx_core.components.guiders import MultiModalGuiderParams
-        from ltx_pipelines.ti2vid_one_stage import TI2VidOneStagePipeline
-
-        pipeline = TI2VidOneStagePipeline(
-            model_paths=ModelPaths.from_monolith(str(args.checkpoint), str(args.gemma_root)),
-            loras=(),
-            offload_mode=offload_mode,
-        )
-        _configure_official_sdpa(pipeline)
-        video, audio, _ = pipeline(
-            prompt=request["prompt"],
-            negative_prompt=request["negative_prompt"],
-            seed=request["seed"],
-            height=request["height"],
-            width=request["width"],
-            num_frames=request["num_frames"],
-            frame_rate=request["fps"],
-            num_inference_steps=request["num_inference_steps"],
-            video_guider_params=MultiModalGuiderParams(
-                cfg_scale=request["video_cfg_scale"],
-                stg_scale=request["video_stg_scale"],
-                rescale_scale=request["video_rescale_scale"],
-                modality_scale=request["video_modality_scale"],
-                skip_step=0,
-                stg_blocks=request["video_stg_blocks"],
-            ),
-            audio_guider_params=MultiModalGuiderParams(
-                cfg_scale=request["audio_cfg_scale"],
-                stg_scale=request["audio_stg_scale"],
-                rescale_scale=request["audio_rescale_scale"],
-                modality_scale=request["audio_modality_scale"],
-                skip_step=0,
-                stg_blocks=request["audio_stg_blocks"],
-            ),
-            images=images,
-            max_batch_size=4,
-        )
-        checkpoint_metadata = {
-            "checkpoint": str(args.checkpoint),
-            "pipeline": "TI2VidOneStagePipeline",
-        }
-
-    video_tensor = torch.cat([chunk.detach().cpu() for chunk in video], dim=0)
-    _save_outputs(
-        args.output_dir,
-        video=_canonical_video(video_tensor),
-        audio=audio.waveform,
-        audio_sample_rate=audio.sampling_rate,
-        fps=float(request["fps"]),
-        metadata={
-            "backend": "official",
-            "attention_backend": "torch_sdpa_cudnn_bf16",
-            "official_revision": os.environ.get("VLLM_TEST_LTX_OFFICIAL_REVISION"),
-            "seed": request["seed"],
-            "sigmas": request.get("sigmas"),
-            "stage_1_sigmas": request.get("stage_1_sigmas"),
-            "stage_2_sigmas": request.get("stage_2_sigmas"),
-            **_accelerator_peak_memory_stats(),
-            **checkpoint_metadata,
-        },
-    )
-
-
-def _omni_worker_peak_memory_mb(output: Any) -> float:
-    result = output[0] if isinstance(output, list) and output else output
-    return float(getattr(result, "peak_memory_mb", 0.0) or 0.0)
-
-
-def _unwrap_omni_output(output: Any) -> tuple[Any, Any, int]:
-    from vllm_omni.outputs import OmniRequestOutput
-
-    audio = None
-    audio_sample_rate = None
-    frames = output[0] if isinstance(output, list) and output else output
-    if isinstance(frames, OmniRequestOutput):
-        multimodal_output = frames.multimodal_output or {}
-        audio = multimodal_output.get("audio")
-        audio_sample_rate = multimodal_output.get("audio_sample_rate")
-        nested_output = getattr(frames, "request_output", None)
-        if frames.is_pipeline_output and isinstance(nested_output, OmniRequestOutput):
-            frames = nested_output
-            multimodal_output = frames.multimodal_output or {}
-            audio = multimodal_output.get("audio", audio)
-            audio_sample_rate = multimodal_output.get("audio_sample_rate", audio_sample_rate)
-        if isinstance(frames, OmniRequestOutput):
-            if not frames.images:
-                raise ValueError("No video frames found in OmniRequestOutput")
-            frames = frames.images
-
-    if isinstance(frames, list) and len(frames) == 1:
-        frames = frames[0]
-    if isinstance(frames, tuple) and len(frames) == 2:
-        frames, audio = frames
-    elif isinstance(frames, dict):
-        audio = frames.get("audio", audio)
-        audio_sample_rate = frames.get("audio_sample_rate", audio_sample_rate)
-        frames = frames.get("frames", frames.get("video"))
-
-    if frames is None or audio is None or audio_sample_rate is None:
-        raise ValueError("Omni output did not contain video, audio, and audio_sample_rate")
-    return frames, audio, int(audio_sample_rate)
-
-
-def _canonical_video(video: Any) -> torch.Tensor:
-    if isinstance(video, list):
-        if len(video) == 1:
-            return _canonical_video(video[0])
-        tensors = [torch.as_tensor(np.asarray(item) if not isinstance(item, torch.Tensor) else item) for item in video]
-        if all(item.ndim == 3 for item in tensors):
-            video = torch.stack(tensors)
-        elif all(item.ndim == 4 for item in tensors):
-            video = torch.cat(tensors)
-        else:
-            raise ValueError(f"Cannot combine video items with dimensions {[item.ndim for item in tensors]}")
-    tensor = torch.as_tensor(np.asarray(video) if not isinstance(video, torch.Tensor) else video).detach().cpu()
-    if tensor.ndim == 5:
-        tensor = tensor[0]
-    if tensor.ndim != 4:
-        raise ValueError(f"Expected a 4D video tensor, got {tuple(tensor.shape)}")
-    if tensor.shape[-1] in (3, 4):
-        tensor = tensor[..., :3]
-    elif tensor.shape[1] in (3, 4):
-        tensor = tensor[:, :3].permute(0, 2, 3, 1)
-    elif tensor.shape[0] in (3, 4):
-        tensor = tensor[:3].permute(1, 2, 3, 0)
-    else:
-        raise ValueError(f"Cannot infer video channel dimension from {tuple(tensor.shape)}")
-    tensor = tensor.float()
-    if tensor.numel() and tensor.max() > 1:
-        tensor = tensor / 255.0
-    elif tensor.numel() and tensor.min() < 0:
-        tensor = tensor.clamp(-1.0, 1.0).add(1.0).mul(0.5)
-    return tensor.clamp(0.0, 1.0)
-
-
-@torch.inference_mode()
-def _run_omni(args: argparse.Namespace, request: dict[str, Any]) -> None:
-    if args.model is None or args.model_class_name is None:
-        raise ValueError("Omni backend requires --model and --model-class-name")
-
-    # vLLM imports can otherwise change the process-wide Gemma dispatch.
-    torch.backends.cuda.enable_cudnn_sdp(False)
-    _reset_accelerator_peak_memory_stats()
-
-    attention_config = {"default": {"backend": "CUDNN_ATTN"}}
-
-    from vllm_omni.diffusion.data import DiffusionParallelConfig
-    from vllm_omni.diffusion.utils.param_utils import apply_declared_extra_args
-    from vllm_omni.entrypoints.omni import Omni
-    from vllm_omni.inputs.data import OmniDiffusionSamplingParams
-    from vllm_omni.model_extras import get_extra_body_params
-    from vllm_omni.platforms import current_omni_platform
-
-    generator = torch.Generator(device=current_omni_platform.device_type).manual_seed(request["seed"])
-    omni = Omni(
-        model=args.model,
-        model_class_name=args.model_class_name,
-        enforce_eager=True,
-        enable_layerwise_offload=args.enable_layerwise_offload,
-        diffusion_attention_config=attention_config,
-        parallel_config=DiffusionParallelConfig(),
-    )
-    try:
-        # The explicit public class argument selects the worker pipeline. The
-        # parent Omni proxy can only expose the checkpoint's base class, which
-        # is insufficient to distinguish LTX execution modes.
-        model_class_name = args.model_class_name
-        sampling_params = OmniDiffusionSamplingParams(
-            height=request["height"],
-            width=request["width"],
-            generator=generator,
-            guidance_scale=None,
-            num_inference_steps=request["num_inference_steps"],
-            num_frames=request["num_frames"],
-            fps=request["fps"],
-            frame_rate=float(request["fps"]),
-            output_type="np",
-        )
-        extra_args = {
-            key: value for key, value in request.items() if key.startswith("video_") or key.startswith("audio_")
-        }
-        for key in ("sigmas", "stage_1_sigmas", "stage_2_sigmas", "image_crf"):
-            if key in request:
-                extra_args[key] = request[key]
-        apply_declared_extra_args(sampling_params, get_extra_body_params(model_class_name), extra_args)
-        prompt: dict[str, Any] = {"prompt": request["prompt"]}
-        if "negative_prompt" in request:
-            prompt["negative_prompt"] = request["negative_prompt"]
-        image_path = request.get("image")
-        if image_path is not None:
-            with Image.open(str(image_path)) as source_image:
-                image = source_image.convert("RGB")
-                image.load()
-            prompt["multi_modal_data"] = {"image": image}
-        output = omni.generate(prompt, sampling_params)
-        video, audio, audio_sample_rate = _unwrap_omni_output(output)
-        audio_tensor = torch.as_tensor(np.asarray(audio) if not isinstance(audio, torch.Tensor) else audio)
-        _save_outputs(
-            args.output_dir,
-            video=_canonical_video(video),
-            audio=audio_tensor,
-            audio_sample_rate=audio_sample_rate,
-            fps=float(request["fps"]),
-            metadata={
-                "backend": "omni",
-                "attention_backend": "torch_sdpa_cudnn_bf16",
-                "seed": request["seed"],
-                "sigmas": request.get("sigmas"),
-                "model": args.model,
-                "stage_1_sigmas": request.get("stage_1_sigmas"),
-                "stage_2_sigmas": request.get("stage_2_sigmas"),
-                "model_class_name": model_class_name,
-                "worker_peak_memory_mb": _omni_worker_peak_memory_mb(output),
-                **_accelerator_peak_memory_stats(),
-            },
-        )
-    finally:
-        omni.shutdown()
-
-
-def main() -> None:
-    args = _parse_args()
-    request = json.loads(args.request.read_text())
-    if args.backend == "official":
-        _run_official(args, request)
-    else:
-        _run_omni(args, request)
-
-
-if __name__ == "__main__":
-    main()
@@ -1,753 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-
-"""E2E accuracy guards against pinned official Lightricks LTX references.
-
-The LTX-2/2.3 comparison runs both runtimes through PyTorch SDPA and uses
-``max_batch_size=4`` in the official reference to match Omni's fused guidance
-batch. Video and audio guidance use the official non-HQ one-stage defaults;
-only the generation shape and step count are reduced for CI runtime.
-LTX-2.5 runs the official split-artifact pipelines with connector weights from
-the same Diffusers checkpoint under test. The distilled case covers the fixed
-two-stage schedule; the Full/SFT case covers raw dev weights and an explicit
-shared one-stage schedule.
-"""
-
-from __future__ import annotations
-
-import json
-import math
-import os
-import shutil
-import subprocess
-import sys
-import time
-from pathlib import Path
-
-import numpy as np
-import pytest
-import torch
-from huggingface_hub import hf_hub_download, snapshot_download
-from torchmetrics.image import PeakSignalNoiseRatio, StructuralSimilarityIndexMeasure
-
-from tests.e2e.accuracy.helpers import reset_artifact_dir
-from tests.helpers import skip_if_gated_repo_inaccessible
-from tests.helpers.mark import hardware_test
-
-OFFICIAL_REPOSITORY = "https://github.com/Lightricks/LTX-2.git"
-LTX25_OFFICIAL_REVISION = "7954dcb0d986bdc36ef272564a9789ade07fcc65"
-LTX25_OMNI_MODEL_ID = "Lightricks/LTX-2.5-Diffusers"
-LTX25_OMNI_MODEL_REVISION = "a6de4b5354f078db24d9cf4778c14846788aea3d"
-LTX25_OFFICIAL_MODEL_ID = "Lightricks/LTX-2.5"
-LTX25_OFFICIAL_MODEL_REVISION = "8a4ff96f581e72bedc1b44367581c49d544a05f1"
-LTX25_OFFICIAL_COMMON_FILES = {
-    "text_encoder": "text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors",
-    "video_vae": "vae/ltx-2.5-video-vae-conv-bf16.safetensors",
-    "audio_vae": "vae/ltx-2.5-audio-vae-bf16.safetensors",
-}
-LTX25_OFFICIAL_VARIANT_FILES = {
-    "distilled": {
-        "transformer": "diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors",
-        "spatial_upsampler": "latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors",
-    },
-    "full": {
-        "transformer": "diffusion_models/ltx-2.5-22b-dev-transformer-bf16.safetensors",
-    },
-    "full_two_stage": {
-        "transformer": "diffusion_models/ltx-2.5-22b-dev-transformer-bf16.safetensors",
-        "spatial_upsampler": "latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors",
-        "distilled_lora": "loras/ltx-2.5-22b-distilled-lora-450-bf16.safetensors",
-    },
-}
-# Version selected by the pinned official source's uv.lock. Keep it isolated
-# from Omni's runtime and development dependencies.
-OFFICIAL_OPENIMAGEIO_VERSION = "3.1.11.0"
-LTX25_PROMPT = "A cinematic shot of a red fox walking through a snowy forest at dawn, the camera tracking alongside."
-LTX25_I2V_IMAGE_REPO = "huggingface/documentation-images"
-LTX25_I2V_IMAGE_FILENAME = "diffusers/svd/rocket.png"
-LTX25_I2V_IMAGE_REVISION = "645d8364f0c7a101180b364811b5a11a362e4010"
-NEGATIVE_PROMPT = (
-    "blurry, out of focus, overexposed, underexposed, low contrast, washed out colors, excessive noise, "
-    "grainy texture, poor lighting, flickering, motion blur, distorted proportions, unnatural skin tones, "
-    "deformed facial features, asymmetrical face, missing facial features, extra limbs, disfigured hands, "
-    "wrong hand count, artifacts around text, inconsistent perspective, camera shake, incorrect depth of field, "
-    "background too sharp, background clutter, distracting reflections, harsh shadows, inconsistent lighting "
-    "direction, color banding, cartoonish rendering, 3D CGI look, unrealistic materials, uncanny valley effect, "
-    "incorrect ethnicity, wrong gender, exaggerated expressions, wrong gaze direction, mismatched lip sync, "
-    "silent or muted audio, distorted voice, robotic voice, echo, background noise, off-sync audio, incorrect "
-    "dialogue, added dialogue, repetitive speech, jittery movement, awkward pauses, incorrect timing, unnatural "
-    "transitions, inconsistent framing, tilted camera, flat lighting, inconsistent tone, cinematic oversaturation, "
-    "stylized filters, or AI artifacts."
-)
-
-# Release parity exercises the accelerated BF16 cuDNN path in both runtimes.
-# Degenerate one-token attention falls back to PyTorch SDPA MATH.
-ATTENTION_BACKEND = "torch_sdpa_cudnn_bf16"
-VIDEO_SSIM_MEAN_THRESHOLD = 0.95
-VIDEO_SSIM_MIN_THRESHOLD = 0.90
-VIDEO_PSNR_MEAN_THRESHOLD = 30.0
-AUDIO_RELATIVE_L2_THRESHOLD = 0.2
-AUDIO_COSINE_THRESHOLD = 0.95
-
-# Decoded-output gates for the pinned official/Omni BF16 cuDNN comparison.
-# The metrics artifact preserves the exact observed values.
-LTX25_VIDEO_SSIM_MEAN_THRESHOLD = 0.99
-LTX25_I2V_VIDEO_SSIM_MEAN_THRESHOLD = 0.99
-LTX25_VIDEO_SSIM_MIN_THRESHOLD = 0.99
-LTX25_VIDEO_PSNR_MEAN_THRESHOLD = 40.0
-LTX25_VIDEO_LPIPS_MEAN_THRESHOLD = 0.01
-LTX25_AUDIO_RELATIVE_L2_THRESHOLD = 0.32
-LTX25_AUDIO_COSINE_THRESHOLD = 0.95
-
-# Full/SFT uses independently converted dev weights, but the release contract
-# still requires decoded-video parity at the same 0.99 SSIM floor as the
-# distilled paths when both runtimes use BF16 cuDNN attention.
-LTX25_FULL_VIDEO_SSIM_MEAN_THRESHOLD = 0.99
-LTX25_FULL_VIDEO_SSIM_MIN_THRESHOLD = 0.99
-LTX25_FULL_VIDEO_PSNR_MEAN_THRESHOLD = 40.0
-LTX25_FULL_AUDIO_RELATIVE_L2_THRESHOLD = 0.3
-LTX25_FULL_VIDEO_LPIPS_MEAN_THRESHOLD = 0.01
-LTX25_FULL_I2V_VIDEO_LPIPS_MEAN_THRESHOLD = 0.02
-LTX25_FULL_AUDIO_COSINE_THRESHOLD = AUDIO_COSINE_THRESHOLD
-LTX25_STAGE_2_SIGMAS = [0.909375, 0.725, 0.421875, 0.0]
-
-
-def test_ltx_reference_runner_unwraps_flattened_pipeline_output() -> None:
-    from vllm_omni.outputs import OmniRequestOutput
-
-    from .run_ltx25_reference import _omni_worker_peak_memory_mb, _unwrap_omni_output
-
-    frame = object()
-    audio = object()
-    output = OmniRequestOutput(
-        stage_id=0,
-        images=[frame],
-        _multimodal_output={"audio": audio, "audio_sample_rate": 48_000},
-    )
-
-    frames, actual_audio, sample_rate = _unwrap_omni_output(output)
-
-    assert frames is frame
-    assert actual_audio is audio
-    assert sample_rate == 48_000
-    output.peak_memory_mb = 12_345.0
-    assert _omni_worker_peak_memory_mb([output]) == 12_345.0
-
-
-def _run(command: list[str], *, env: dict[str, str], timeout: int = 1800) -> None:
-    start = time.perf_counter()
-    subprocess.run(command, env=env, timeout=timeout, check=True)
-    print(f"{' '.join(command[:3])} finished in {time.perf_counter() - start:.1f}s")
-
-
-def _clone_pinned_source(root: Path, repository: str, revision: str) -> None:
-    root.mkdir(parents=True)
-    subprocess.run(["git", "init", "-q", str(root)], check=True)
-    subprocess.run(["git", "-C", str(root), "remote", "add", "origin", repository], check=True)
-    last_error: subprocess.CalledProcessError | None = None
-    for attempt in range(3):
-        try:
-            subprocess.run(
-                ["git", "-C", str(root), "fetch", "--depth", "1", "origin", revision],
-                check=True,
-            )
-            last_error = None
-            break
-        except subprocess.CalledProcessError as error:
-            last_error = error
-            if attempt < 2:
-                time.sleep(5 * (attempt + 1))
-    if last_error is not None:
-        raise last_error
-    subprocess.run(["git", "-C", str(root), "checkout", "-q", "--detach", "FETCH_HEAD"], check=True)
-
-
-def _git_revision(root: Path) -> str | None:
-    try:
-        return subprocess.check_output(
-            ["git", "-C", str(root), "rev-parse", "HEAD"],
-            stderr=subprocess.DEVNULL,
-            text=True,
-        ).strip()
-    except subprocess.CalledProcessError:
-        return None
-
-
-def _ltx25_official_source(artifact_root: Path) -> tuple[Path, str]:
-    repository = os.environ.get("VLLM_TEST_LTX25_OFFICIAL_REPOSITORY", OFFICIAL_REPOSITORY)
-    revision = os.environ.get("VLLM_TEST_LTX25_OFFICIAL_REVISION", LTX25_OFFICIAL_REVISION)
-    configured_root = os.environ.get("VLLM_TEST_LTX25_OFFICIAL_ROOT")
-    root = Path(configured_root) if configured_root else artifact_root / f"official-source-{revision[:12]}"
-    actual_revision = _git_revision(root) if root.exists() else None
-    if actual_revision != revision and configured_root:
-        raise AssertionError(f"Official LTX-2.5 source revision mismatch: {actual_revision} != {revision}")
-    if actual_revision != revision:
-        if root.exists():
-            shutil.rmtree(root)
-        _clone_pinned_source(root, repository, revision)
-        actual_revision = _git_revision(root)
-    assert actual_revision == revision, f"Official LTX-2.5 source revision mismatch: {actual_revision} != {revision}"
-    return root, revision
-
-
-def _official_runner_prefix() -> list[str]:
-    """Run the reference with its missing binary dependency isolated from CI."""
-    uv = shutil.which("uv")
-    assert uv is not None, "uv is required to run the pinned official LTX reference"
-    return [
-        uv,
-        "run",
-        "--no-project",
-        "--with",
-        f"openimageio=={OFFICIAL_OPENIMAGEIO_VERSION}",
-        "--python",
-        sys.executable,
-        "python",
-    ]
-
-
-def _resolve_ltx25_omni_model(*, transformer_subfolder: str = "transformer") -> tuple[Path, str, str | None]:
-    if transformer_subfolder not in {"transformer", "transformer_full"}:
-        raise ValueError(f"Unsupported LTX-2.5 transformer subfolder: {transformer_subfolder!r}")
-    configured_model = os.environ.get("VLLM_TEST_LTX25_MODEL")
-    configured_revision = os.environ.get("VLLM_TEST_LTX25_MODEL_REVISION")
-    if configured_model and Path(configured_model).exists():
-        model = Path(configured_model).resolve()
-        snapshot_revision = model.name if model.parent.name == "snapshots" else None
-        assert (model / transformer_subfolder).is_dir(), f"LTX-2.5 component not found: {model / transformer_subfolder}"
-        return model, configured_model, configured_revision or snapshot_revision
-    model_id = configured_model or LTX25_OMNI_MODEL_ID
-    revision = configured_revision
-    if revision is None and model_id == LTX25_OMNI_MODEL_ID:
-        revision = LTX25_OMNI_MODEL_REVISION
-    model = Path(
-        snapshot_download(
-            repo_id=model_id,
-            revision=revision,
-            allow_patterns=[
-                "model_index.json",
-                "audio_vae/*",
-                "connectors/config.json",
-                "connectors/diffusion_pytorch_model.safetensors.index.json",
-                "connectors/diffusion_pytorch_model-*.safetensors",
-                "scheduler/*",
-                "text_encoder/config.json",
-                "text_encoder/generation_config.json",
-                "text_encoder/model*",
-                "tokenizer/*",
-                f"{transformer_subfolder}/*",
-                "vae/*",
-                "vocoder/*",
-            ],
-        )
-    ).resolve()
-    snapshot_revision = model.name if model.parent.name == "snapshots" else None
-    return model, model_id, snapshot_revision or revision
-
-
-def _resolve_ltx25_official_artifacts(
-    *, checkpoint_variant: str = "distilled"
-) -> tuple[dict[str, Path], str, str | None]:
-    try:
-        variant_files = LTX25_OFFICIAL_VARIANT_FILES[checkpoint_variant]
-    except KeyError as exc:
-        raise ValueError(f"Unsupported official LTX-2.5 checkpoint variant: {checkpoint_variant!r}") from exc
-    files = {**LTX25_OFFICIAL_COMMON_FILES, **variant_files}
-    configured_model = os.environ.get("VLLM_TEST_LTX25_OFFICIAL_MODEL")
-    configured_revision = os.environ.get("VLLM_TEST_LTX25_OFFICIAL_MODEL_REVISION")
-    if configured_model and Path(configured_model).exists():
-        root = Path(configured_model).resolve()
-        model_source = configured_model
-        snapshot_revision = root.name if root.parent.name == "snapshots" else None
-        resolved_revision = configured_revision or snapshot_revision
-    else:
-        model_source = configured_model or LTX25_OFFICIAL_MODEL_ID
-        revision = configured_revision
-        if revision is None and model_source == LTX25_OFFICIAL_MODEL_ID:
-            revision = LTX25_OFFICIAL_MODEL_REVISION
-        root = Path(
-            snapshot_download(
-                repo_id=model_source,
-                revision=revision,
-                allow_patterns=list(files.values()),
-            )
-        ).resolve()
-        snapshot_revision = root.name if root.parent.name == "snapshots" else None
-        resolved_revision = snapshot_revision or revision
-
-    artifacts = {name: root / relative_path for name, relative_path in files.items()}
-    missing = [f"{name}: {path}" for name, path in artifacts.items() if not path.is_file()]
-    assert not missing, f"Official LTX-2.5 artifacts are missing: {', '.join(missing)}"
-    return artifacts, model_source, resolved_revision
-
-
-def _resolve_gemma_root(model: Path) -> Path:
-    configured_root = os.environ.get("VLLM_TEST_LTX_GEMMA_ROOT")
-    if configured_root:
-        root = Path(configured_root)
-        assert root.is_dir(), f"Gemma root not found: {root}"
-        return root
-    return model
-
-
-def _resolve_ltx25_image() -> Path:
-    configured_image = os.environ.get("VLLM_TEST_LTX25_I2V_IMAGE")
-    if configured_image:
-        image = Path(configured_image)
-        assert image.is_file(), f"LTX-2.5 I2V conditioning image not found: {image}"
-        return image
-    return Path(
-        hf_hub_download(
-            repo_id=LTX25_I2V_IMAGE_REPO,
-            repo_type="dataset",
-            filename=LTX25_I2V_IMAGE_FILENAME,
-            revision=LTX25_I2V_IMAGE_REVISION,
-        )
-    )
-
-
-def _require_ltx25_model_access() -> None:
-    configured_omni_model = os.environ.get("VLLM_TEST_LTX25_MODEL")
-    if not (configured_omni_model and Path(configured_omni_model).exists()):
-        omni_model = configured_omni_model or LTX25_OMNI_MODEL_ID
-        omni_revision = os.environ.get("VLLM_TEST_LTX25_MODEL_REVISION")
-        if omni_revision is None and omni_model == LTX25_OMNI_MODEL_ID:
-            omni_revision = LTX25_OMNI_MODEL_REVISION
-        skip_if_gated_repo_inaccessible(
-            omni_model,
-            revision=omni_revision,
-            filename="model_index.json",
-        )
-
-    configured_official_model = os.environ.get("VLLM_TEST_LTX25_OFFICIAL_MODEL")
-    if not (configured_official_model and Path(configured_official_model).exists()):
-        official_model = configured_official_model or LTX25_OFFICIAL_MODEL_ID
-        official_revision = os.environ.get("VLLM_TEST_LTX25_OFFICIAL_MODEL_REVISION")
-        if official_revision is None and official_model == LTX25_OFFICIAL_MODEL_ID:
-            official_revision = LTX25_OFFICIAL_MODEL_REVISION
-        skip_if_gated_repo_inaccessible(
-            official_model,
-            revision=official_revision,
-            filename=LTX25_OFFICIAL_COMMON_FILES["text_encoder"],
-        )
-
-
-def _ltx25_request(image: Path | None = None) -> dict[str, object]:
-    request: dict[str, object] = {
-        "prompt": LTX25_PROMPT,
-        # Keep the nightly parity test small enough to run the pinned official
-        # pipeline and Omni sequentially on one H100. The public recipe covers
-        # the release-resolution 1920x1088x121 configuration.
-        "width": 1024,
-        "height": 576,
-        "num_frames": 25,
-        "fps": 24,
-        "num_inference_steps": 8,
-        "seed": 42,
-    }
-    if image is not None:
-        request.update(image=str(image.resolve()), image_crf=18)
-    return request
-
-
-def _ltx25_full_sigmas(num_inference_steps: int) -> list[float]:
-    """Materialize the official LTX-2.5 one-stage schedule in FP32."""
-    sigmas = torch.linspace(1.0, 0.0, num_inference_steps + 1)
-    base_anchor = 1024
-    max_anchor = 4096
-    base_shift = 0.95
-    max_shift = 2.05
-    slope = (max_shift - base_shift) / (max_anchor - base_anchor)
-    sigma_shift = max_anchor * slope + (base_shift - slope * base_anchor)
-    exp_shift = math.exp(sigma_shift)
-    sigmas = torch.where(sigmas != 0, exp_shift / (exp_shift + (1 / sigmas - 1)), 0)
-
-    non_zero = sigmas != 0
-    one_minus_sigmas = 1.0 - sigmas[non_zero]
-    scale = one_minus_sigmas[-1] / (1.0 - 0.1)
-    sigmas[non_zero] = 1.0 - one_minus_sigmas / scale
-    return [float(sigma) for sigma in sigmas]
-
-
-def _ltx25_full_request(image: Path | None = None) -> dict[str, object]:
-    num_inference_steps = 30
-    request: dict[str, object] = {
-        "prompt": LTX25_PROMPT,
-        "negative_prompt": NEGATIVE_PROMPT,
-        # Full guidance expands to a four-way batch. Keep the decoded output
-        # small enough for the pinned official and Omni runs to share one H100.
-        "width": 512,
-        "height": 384,
-        "num_frames": 25,
-        "fps": 24,
-        "num_inference_steps": num_inference_steps,
-        "seed": 42,
-        "video_cfg_scale": 3.0,
-        "audio_cfg_scale": 7.0,
-        "video_stg_scale": 1.0,
-        "audio_stg_scale": 1.0,
-        "video_modality_scale": 3.0,
-        "audio_modality_scale": 3.0,
-        "video_rescale_scale": 0.7,
-        "audio_rescale_scale": 0.7,
-        "video_stg_blocks": [28],
-        "audio_stg_blocks": [28],
-        "sigmas": _ltx25_full_sigmas(num_inference_steps),
-    }
-    if image is not None:
-        request.update(image=str(image.resolve()), image_crf=18)
-    return request
-
-
-def test_ltx25_full_request_pins_official_schedule() -> None:
-    request = _ltx25_full_request()
-    sigmas = request["sigmas"]
-
-    assert isinstance(sigmas, list)
-    assert isinstance(request["num_inference_steps"], int)
-    assert len(sigmas) == request["num_inference_steps"] + 1
-    assert math.isclose(sigmas[0], 1.0, rel_tol=0.0, abs_tol=1e-6)
-    assert sigmas[-1] == 0.0
-    assert all(sigmas[index] > sigmas[index + 1] for index in range(len(sigmas) - 1))
-
-
-def test_ltx25_i2v_requests_pin_official_crf() -> None:
-    image = Path("conditioning.png")
-    for request in (_ltx25_request(image), _ltx25_full_request(image)):
-        assert request["image"] == str(image.resolve())
-        assert request["image_crf"] == 18
-
-
-def _video_metrics(reference: np.ndarray, prediction: np.ndarray) -> dict[str, float]:
-    from benchmarks.diffusion.quantization_quality import compute_lpips_video
-
-    assert reference.shape == prediction.shape
-    assert reference.ndim == 4 and reference.shape[-1] == 3
-    ssim_scores: list[float] = []
-    psnr_scores: list[float] = []
-    for reference_frame, prediction_frame in zip(reference, prediction, strict=True):
-        reference_tensor = torch.from_numpy(reference_frame).permute(2, 0, 1).unsqueeze(0)
-        prediction_tensor = torch.from_numpy(prediction_frame).permute(2, 0, 1).unsqueeze(0)
-        ssim_scores.append(float(StructuralSimilarityIndexMeasure(data_range=1.0)(prediction_tensor, reference_tensor)))
-        psnr_scores.append(float(PeakSignalNoiseRatio(data_range=1.0)(prediction_tensor, reference_tensor)))
-    difference = np.abs(reference.astype(np.float64) - prediction.astype(np.float64))
-    return {
-        "ssim_mean": float(np.mean(ssim_scores)),
-        "ssim_min": float(np.min(ssim_scores)),
-        "psnr_mean_db": float(np.mean(psnr_scores)),
-        "lpips_alex_mean": compute_lpips_video(reference, prediction),
-        "max_abs": float(difference.max()),
-        "mean_abs": float(difference.mean()),
-    }
-
-
-def _canonical_audio(audio: np.ndarray) -> np.ndarray:
-    while audio.ndim > 2 and audio.shape[0] == 1:
-        audio = audio[0]
-    return audio.astype(np.float64)
-
-
-def _audio_metrics(reference: np.ndarray, prediction: np.ndarray) -> dict[str, float | bool]:
-    reference = _canonical_audio(reference)
-    prediction = _canonical_audio(prediction)
-    assert reference.shape == prediction.shape
-    difference = reference - prediction
-    reference_norm = max(float(np.linalg.norm(reference)), 1e-12)
-    prediction_norm = max(float(np.linalg.norm(prediction)), 1e-12)
-    return {
-        "bitwise_equal": bool(np.array_equal(reference, prediction)),
-        "max_abs": float(np.abs(difference).max()),
-        "mean_abs": float(np.abs(difference).mean()),
-        "relative_l2": float(np.linalg.norm(difference) / reference_norm),
-        "cosine_similarity": float(np.vdot(reference.ravel(), prediction.ravel()) / (reference_norm * prediction_norm)),
-    }
-
-
-@pytest.mark.full_model
-@pytest.mark.benchmark
-@pytest.mark.diffusion
-@hardware_test(res={"cuda": "H100"}, num_cards=1)
-@pytest.mark.parametrize("task", ["t2v", "i2v"])
-def test_ltx25_distilled_two_stage_matches_official(accuracy_artifact_root: Path, task: str) -> None:
-    """Compare Omni with the official runtime using the same checkpoint connector weights."""
-    _require_ltx25_model_access()
-    artifact_parent = accuracy_artifact_root / "ltx_official"
-    output_root = reset_artifact_dir(artifact_parent / f"ltx2_5_distilled_{task}")
-    official_root, official_revision = _ltx25_official_source(artifact_parent)
-    official_artifacts, official_model_source, official_model_revision = _resolve_ltx25_official_artifacts()
-    omni_model, omni_model_source, omni_model_revision = _resolve_ltx25_omni_model()
-    image = _resolve_ltx25_image() if task == "i2v" else None
-    request_path = output_root / "request.json"
-    request_path.write_text(json.dumps(_ltx25_request(image), indent=2) + "\n")
-
-    runner = Path(__file__).with_name("run_ltx25_reference.py")
-    runner_args = [
-        str(runner),
-        "--request",
-        str(request_path),
-        "--enable-layerwise-offload",
-    ]
-    env = os.environ.copy()
-    env["VLLM_TEST_LTX_OFFICIAL_REVISION"] = official_revision
-    env["PYTHONUNBUFFERED"] = "1"
-    repository_root = Path(__file__).resolve().parents[4]
-    existing_pythonpath = env.get("PYTHONPATH")
-    env["PYTHONPATH"] = (
-        str(repository_root) if not existing_pythonpath else f"{repository_root}{os.pathsep}{existing_pythonpath}"
-    )
-
-    official_output = output_root / "official"
-    _run(
-        _official_runner_prefix()
-        + runner_args
-        + [
-            "--backend",
-            "official",
-            "--output-dir",
-            str(official_output),
-            "--official-root",
-            str(official_root),
-            "--transformer-path",
-            str(official_artifacts["transformer"]),
-            "--text-encoder-path",
-            str(official_artifacts["text_encoder"]),
-            "--video-vae-path",
-            str(official_artifacts["video_vae"]),
-            "--audio-vae-path",
-            str(official_artifacts["audio_vae"]),
-            "--spatial-upsampler-path",
-            str(official_artifacts["spatial_upsampler"]),
-            "--connector-model",
-            str(omni_model),
-        ],
-        env=env,
-    )
-
-    omni_output = output_root / "omni"
-    _run(
-        [sys.executable]
-        + runner_args
-        + [
-            "--backend",
-            "omni",
-            "--output-dir",
-            str(omni_output),
-            "--model",
-            str(omni_model),
-            "--model-class-name",
-            "LTX2DistilledTwoStagePipeline",
-        ],
-        env=env,
-    )
-
-    official_metadata = json.loads((official_output / "metadata.json").read_text())
-    omni_metadata = json.loads((omni_output / "metadata.json").read_text())
-    assert official_metadata["official_revision"] == official_revision
-    assert official_metadata["pipeline"] == "DistilledPipeline"
-    assert official_metadata["attention_backend"] == ATTENTION_BACKEND
-    assert official_metadata["connector_model"] == str(omni_model)
-    assert omni_metadata["model_class_name"] == "LTX2DistilledTwoStagePipeline"
-    assert omni_metadata["attention_backend"] == ATTENTION_BACKEND
-    assert official_metadata["audio_sample_rate"] == omni_metadata["audio_sample_rate"]
-    video_metrics = _video_metrics(
-        np.load(official_output / "video.npy"),
-        np.load(omni_output / "video.npy"),
-    )
-    audio_metrics = _audio_metrics(
-        np.load(official_output / "audio.npy"),
-        np.load(omni_output / "audio.npy"),
-    )
-    result = {
-        "case": "ltx2_5_distilled",
-        "task": task,
-        "attention_backend": ATTENTION_BACKEND,
-        "official_revision": official_revision,
-        "official_model": official_model_source,
-        "official_model_revision": official_model_revision,
-        "official_connector_model": omni_model_source,
-        "official_connector_revision": omni_model_revision,
-        "omni_model": omni_model_source,
-        "omni_model_revision": omni_model_revision,
-        "resolved_omni_model_path": str(omni_model),
-        "peak_memory_mb": {
-            "official_allocated": official_metadata["peak_memory_allocated_mb"],
-            "official_reserved": official_metadata["peak_memory_reserved_mb"],
-            "omni_worker_reported": omni_metadata["worker_peak_memory_mb"],
-            "omni_parent_allocated": omni_metadata["peak_memory_allocated_mb"],
-            "omni_parent_reserved": omni_metadata["peak_memory_reserved_mb"],
-        },
-        "video": video_metrics,
-        "audio": audio_metrics,
-    }
-    (output_root / "metrics.json").write_text(json.dumps(result, indent=2) + "\n")
-    print(json.dumps(result, indent=2))
-
-    ssim_mean_threshold = LTX25_VIDEO_SSIM_MEAN_THRESHOLD if task == "t2v" else LTX25_I2V_VIDEO_SSIM_MEAN_THRESHOLD
-    assert video_metrics["ssim_mean"] >= ssim_mean_threshold
-    assert video_metrics["ssim_min"] >= LTX25_VIDEO_SSIM_MIN_THRESHOLD
-    assert video_metrics["psnr_mean_db"] >= LTX25_VIDEO_PSNR_MEAN_THRESHOLD
-    assert video_metrics["lpips_alex_mean"] <= LTX25_VIDEO_LPIPS_MEAN_THRESHOLD
-    assert audio_metrics["cosine_similarity"] >= LTX25_AUDIO_COSINE_THRESHOLD
-    assert audio_metrics["relative_l2"] <= LTX25_AUDIO_RELATIVE_L2_THRESHOLD
-
-
-@pytest.mark.full_model
-@pytest.mark.benchmark
-@pytest.mark.diffusion
-@hardware_test(res={"cuda": "H100"}, num_cards=1)
-@pytest.mark.parametrize("task", ["t2v", "i2v"])
-@pytest.mark.parametrize("pipeline_mode", ["full_one_stage", "full_two_stage"])
-def test_ltx25_full_matches_official(accuracy_artifact_root: Path, task: str, pipeline_mode: str) -> None:
-    """Compare raw official Full/SFT weights with Omni's converted Full pipeline."""
-    _require_ltx25_model_access()
-    artifact_parent = accuracy_artifact_root / "ltx_official"
-    output_root = reset_artifact_dir(artifact_parent / f"ltx2_5_{pipeline_mode}_{task}")
-    model_class_name = {
-        "full_one_stage": "LTX2Pipeline",
-        "full_two_stage": "LTX2TwoStagePipeline",
-    }[pipeline_mode]
-    official_root, official_revision = _ltx25_official_source(artifact_parent)
-    official_artifacts, official_model_source, official_model_revision = _resolve_ltx25_official_artifacts(
-        checkpoint_variant="full" if pipeline_mode == "full_one_stage" else "full_two_stage"
-    )
-    omni_model, omni_model_source, omni_model_revision = _resolve_ltx25_omni_model(
-        transformer_subfolder="transformer_full"
-    )
-    image = _resolve_ltx25_image() if task == "i2v" else None
-    request = _ltx25_full_request(image)
-    if pipeline_mode == "full_two_stage":
-        request["stage_1_sigmas"] = request.pop("sigmas")
-        request["stage_2_sigmas"] = LTX25_STAGE_2_SIGMAS
-    request_path = output_root / "request.json"
-    request_path.write_text(json.dumps(request, indent=2) + "\n")
-
-    runner = Path(__file__).with_name("run_ltx25_reference.py")
-    runner_args = [
-        str(runner),
-        "--request",
-        str(request_path),
-        "--enable-layerwise-offload",
-    ]
-    env = os.environ.copy()
-    env["VLLM_TEST_LTX_OFFICIAL_REVISION"] = official_revision
-    env["PYTHONUNBUFFERED"] = "1"
-    repository_root = Path(__file__).resolve().parents[4]
-    existing_pythonpath = env.get("PYTHONPATH")
-    env["PYTHONPATH"] = (
-        str(repository_root) if not existing_pythonpath else f"{repository_root}{os.pathsep}{existing_pythonpath}"
-    )
-
-    official_output = output_root / "official"
-    official_sidecar_args: list[str] = []
-    if pipeline_mode == "full_two_stage":
-        official_sidecar_args = [
-            "--spatial-upsampler-path",
-            str(official_artifacts["spatial_upsampler"]),
-            "--distilled-lora-path",
-            str(official_artifacts["distilled_lora"]),
-        ]
-
-    _run(
-        _official_runner_prefix()
-        + runner_args
-        + [
-            "--backend",
-            "official",
-            "--output-dir",
-            str(official_output),
-            "--official-root",
-            str(official_root),
-            "--official-pipeline",
-            pipeline_mode,
-            "--transformer-path",
-            str(official_artifacts["transformer"]),
-            "--text-encoder-path",
-            str(official_artifacts["text_encoder"]),
-            "--video-vae-path",
-            str(official_artifacts["video_vae"]),
-            "--audio-vae-path",
-            str(official_artifacts["audio_vae"]),
-            "--connector-model",
-            str(omni_model),
-        ]
-        + official_sidecar_args,
-        env=env,
-    )
-
-    omni_output = output_root / "omni"
-    _run(
-        [sys.executable]
-        + runner_args
-        + [
-            "--backend",
-            "omni",
-            "--output-dir",
-            str(omni_output),
-            "--model",
-            str(omni_model),
-            "--model-class-name",
-            model_class_name,
-        ],
-        env=env,
-    )
-
-    official_metadata = json.loads((official_output / "metadata.json").read_text())
-    omni_metadata = json.loads((omni_output / "metadata.json").read_text())
-    assert official_metadata["official_revision"] == official_revision
-    assert (
-        official_metadata["pipeline"]
-        == {
-            "full_one_stage": "TI2VidOneStagePipeline",
-            "full_two_stage": "TI2VidTwoStagesPipeline",
-        }[pipeline_mode]
-    )
-    assert official_metadata["attention_backend"] == ATTENTION_BACKEND
-    assert official_metadata["connector_model"] == str(omni_model)
-    assert omni_metadata["model_class_name"] == model_class_name
-    assert omni_metadata["attention_backend"] == ATTENTION_BACKEND
-    assert official_metadata["seed"] == omni_metadata["seed"]
-    for schedule_name in ("sigmas", "stage_1_sigmas", "stage_2_sigmas"):
-        assert official_metadata[schedule_name] == omni_metadata[schedule_name]
-    assert official_metadata["audio_sample_rate"] == omni_metadata["audio_sample_rate"]
-    video_metrics = _video_metrics(
-        np.load(official_output / "video.npy"),
-        np.load(omni_output / "video.npy"),
-    )
-    audio_metrics = _audio_metrics(
-        np.load(official_output / "audio.npy"),
-        np.load(omni_output / "audio.npy"),
-    )
-    result = {
-        "case": f"ltx2_5_{pipeline_mode}",
-        "task": task,
-        "attention_backend": ATTENTION_BACKEND,
-        "official_revision": official_revision,
-        "official_model": official_model_source,
-        "official_model_revision": official_model_revision,
-        "official_connector_model": omni_model_source,
-        "official_connector_revision": omni_model_revision,
-        "omni_model": omni_model_source,
-        "omni_model_revision": omni_model_revision,
-        "resolved_omni_model_path": str(omni_model),
-        "peak_memory_mb": {
-            "official_allocated": official_metadata["peak_memory_allocated_mb"],
-            "official_reserved": official_metadata["peak_memory_reserved_mb"],
-            "omni_worker_reported": omni_metadata["worker_peak_memory_mb"],
-            "omni_parent_allocated": omni_metadata["peak_memory_allocated_mb"],
-            "omni_parent_reserved": omni_metadata["peak_memory_reserved_mb"],
-        },
-        "video": video_metrics,
-        "audio": audio_metrics,
-    }
-    (output_root / "metrics.json").write_text(json.dumps(result, indent=2) + "\n")
-    print(json.dumps(result, indent=2))
-
-    assert video_metrics["ssim_mean"] >= LTX25_FULL_VIDEO_SSIM_MEAN_THRESHOLD
-    assert video_metrics["ssim_min"] >= LTX25_FULL_VIDEO_SSIM_MIN_THRESHOLD
-    assert video_metrics["psnr_mean_db"] >= LTX25_FULL_VIDEO_PSNR_MEAN_THRESHOLD
-    lpips_threshold = (
-        LTX25_FULL_VIDEO_LPIPS_MEAN_THRESHOLD if task == "t2v" else LTX25_FULL_I2V_VIDEO_LPIPS_MEAN_THRESHOLD
-    )
-    assert audio_metrics["relative_l2"] <= LTX25_FULL_AUDIO_RELATIVE_L2_THRESHOLD
-    assert video_metrics["lpips_alex_mean"] <= lpips_threshold
-    assert audio_metrics["cosine_similarity"] >= LTX25_FULL_AUDIO_COSINE_THRESHOLD
@@ -29,14 +29,14 @@ HEIGHT = 512
 NUM_INFERENCE_STEPS = 20
 TRUE_CFG_SCALE = 4.0
 SEED = 42
-SSIM_THRESHOLD = 0.94
-# With --fa-deterministic, H100 nightlies stably land near SSIM 0.958 / PSNR 27.8
-# vs Diffusers (see #5734 / #5963), so the primary gates are 0.94 / 27.0.
-# The SDPA-matched fallback (hub kernel unavailable for the image's torch)
-# landed at SSIM 0.9687 in build 2954 — keep a slightly tighter SSIM gate for
-# that path only.
+SSIM_THRESHOLD = 0.97
+# The FA3-hub-matched comparison clears 0.97 with margin. The SDPA-matched
+# fallback (hub kernel unavailable for the image's torch) landed at 0.9687 in
+# build 2954 — the attention math matches but the surrounding stacks (vLLM
+# compiled path vs eager diffusers) still differ slightly. Keep the strict
+# gate for the FA3 path and a slightly looser one for the fallback only.
 SSIM_THRESHOLD_SDPA_FALLBACK = 0.96
-PSNR_THRESHOLD = 27.0
+PSNR_THRESHOLD = 30.0

 MODEL_2512_ID = "Qwen/Qwen-Image-2512"
 MODEL_2512_ENV_VAR = "QWEN_IMAGE_2512_MODEL"
@@ -46,7 +46,7 @@ REFERENCE_TEXT_IMG2TEXT = (
 _OMNI_RUNNER_PARAM = (MODEL_NAME, STAGE_CONFIG)

 pytestmark = [
-    pytest.mark.slow,
+    pytest.mark.advanced_model,
    pytest.mark.diffusion,
    pytest.mark.parametrize("omni_runner", [_OMNI_RUNNER_PARAM], indirect=True),
 ]
@@ -72,7 +72,7 @@ def _send_inpaint_with_generator(
    )


-@pytest.mark.slow
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@hardware_test(res={"cuda": "L4"}, num_cards=1)
 def test_flux2_klein_inpaint_deterministic(omni_runner_handler: OmniRunnerHandler):
@@ -93,7 +93,7 @@ def test_flux2_klein_inpaint_deterministic(omni_runner_handler: OmniRunnerHandle
    )


-@pytest.mark.slow
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@hardware_test(res={"cuda": "L4"}, num_cards=1)
 def test_flux2_klein_inpaint_different_seeds_different_output(omni_runner_handler: OmniRunnerHandler):
@@ -1,345 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""E2E offline inference tests for the Gepard-1.0 single-stage native-AR TTS.
-
-Zero-shot only (the model's default learned voice). These tests need a GPU and
-the NeMo NanoCodec, so the whole file runs in the weekly tier -- see the comment
-above ``pytestmark``. The window arithmetic behind the streaming decode is
-covered on CPU in ``tests/model_executor/models/test_gepard_window.py``.
-"""
-
-from __future__ import annotations
-
-import functools
-import tempfile
-from pathlib import Path
-from typing import NamedTuple
-
-import pytest
-import soundfile as sf
-import torch
-from transformers import AutoTokenizer
-
-from tests.helpers.mark import hardware_test
-from tests.helpers.media import convert_audio_file_to_text
-from tests.helpers.runtime import OmniRunner  # noqa: F401  (indirect fixture)
-from tests.helpers.stage_config import get_deploy_config_path
-from vllm_omni import Omni
-from vllm_omni.model_executor.models.gepard.configuration_gepard import GepardConfig
-from vllm_omni.model_executor.models.gepard.prompt import build_gepard_prompt_ids
-
-# The codec decoder is an optional dependency; skip rather than fail the engine
-# inside the worker process when it is absent. The CI image does not carry it,
-# so the weekly step installs it first -- see .buildkite/cuda/test-weekly.yml.
-# Install recipe: examples/offline_inference/text_to_speech/README.md.
-pytest.importorskip("nemo.collections.tts.models")
-
-MODEL_NAME = "nineninesix/gepard-1.0"
-STAGE_CONFIG = get_deploy_config_path("gepard.yaml")
-SAMPLE_RATE = 22050
-# NanoCodec upsamples one FSQ frame to exactly this many samples.
-SAMPLES_PER_FRAME = 1024
-
-
-@functools.cache
-def _config() -> GepardConfig:
-    """The checkpoint's own layout, not the class defaults.
-
-    The worker builds its config from the ``gepard_config.json`` sidecar, so a
-    prompt assembled from the defaults would silently use a different layout
-    the day the sidecar moves — and a mismatched layout degrades the speech
-    rather than failing, which is the hardest way for this to be found.
-    """
-    return GepardConfig.from_checkpoint(MODEL_NAME)
-
-
-# Prompt -> a word that appears in no other prompt, so a clip can be traced
-# back to the request that asked for it.
-#
-# These are not interchangeable. Batching perturbs every request's sampling
-# trajectory, and on some trajectories this checkpoint's stop head never fires;
-# this set was screened over 10 batched runs with no runaway. Re-screen the
-# whole set when changing one, not just the row changed.
-_DISTINGUISHABLE_PROMPTS = {
-    "He drinks coffee every morning.": "coffee",
-    "Machine learning is interesting.": "learning",
-    "Please close the window before leaving.": "window",
-    "My favorite color is purple.": "purple",
-}
-
-# Above text_repetition.apply_below, so the layout is voiced once instead of
-# repeated. Every other prompt here takes the repeated branch.
-_LONG_TEXT = (
-    "The morning train was late again, so I walked the last two miles along the "
-    "river and reached the office just before nine."
-)
-
-_OMNI_RUNNER_PARAM = (
-    MODEL_NAME,
-    STAGE_CONFIG,
-    {"trust_remote_code": True},
-)
-
-# Only a common word is checked, which needs less ASR than the shared helper's
-# ``small`` default; a miss escalates to it.
-_ASR_MODEL = "base"
-_ASR_ESCALATION_MODEL = "small"
-
-# Transcription dominates this file's runtime -- Whisper runs on CPU on a
-# single-GPU runner -- so only two of the four concurrent clips are read as
-# speech. The pairwise waveform comparison covers all four.
-_TRANSCRIBED_CLIPS = 2
-
-# No level marker at module scope: each test carries the one matching how often
-# its claim needs re-checking, though the weekly step collects the file whole.
-# There is deliberately no core_model row -- under dummy weights the stop head is
-# random, so every request would run to gepard.yaml's max_tokens.
-pytestmark = [
-    pytest.mark.tts,
-    pytest.mark.parametrize("omni_runner", [_OMNI_RUNNER_PARAM], indirect=True),
-]
-
-
-def _build_request(text: str) -> dict:
-    """Assemble the [speaker slots, SOT, text, EOT, SOS] prompt."""
-    tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, trust_remote_code=True)
-    prompt_token_ids = build_gepard_prompt_ids(
-        tokenizer(text, add_special_tokens=False)["input_ids"],
-        config=_config(),
-    )
-    return {
-        "prompt_token_ids": prompt_token_ids,
-        "additional_information": {"text": [text]},
-    }
-
-
-def _extract_audio(mm) -> torch.Tensor | None:
-    # The consolidation path renames "model_outputs" to "audio".
-    audio = mm.get("audio")
-    if audio is None:
-        audio = mm.get("model_outputs")
-    if isinstance(audio, list):
-        audio = audio[0] if len(audio) else None
-    return audio
-
-
-class _Clip(NamedTuple):
-    wav: torch.Tensor
-    # Counts the frames the codec was handed using only what the engine
-    # returned, never the model's internals.
-    committed_frames: int
-
-
-def _synthesize_all(omni: Omni, texts: list[str]) -> list[_Clip]:
-    """Submit every text in one call; return one clip per request, in
-    submission order.
-
-    ``Omni.generate`` yields in completion order and never sorts, so the pairing
-    is restored from the request id prefix it assigns (``f"{i}_{uuid4()}"``).
-    Without that, the only testable claim is that the batch as a whole looks
-    plausible, which a stable swap between two requests satisfies.
-
-    Nothing here authors a SamplingParams: a fresh one would replace the stage
-    defaults rather than merge over them, dropping the pipeline's stop token.
-    """
-    outputs = omni.generate([_build_request(t) for t in texts])
-    assert len(outputs) == len(texts), f"expected {len(texts)} outputs, got {len(outputs)}"
-
-    by_index: dict[int, _Clip] = {}
-    for stage_outputs in outputs:
-        index = int(str(stage_outputs.request_id).split("_", 1)[0])
-        waveform = None
-        frames = 0
-        for out in stage_outputs.outputs:
-            # Explicit None check: `x or []` would ask an array for a truth value.
-            token_ids = out.token_ids if out.token_ids is not None else []
-            # STOP sits just past head0's range, so a token equal to it is
-            # never a committed frame.
-            frames += sum(1 for t in token_ids if int(t) != _config().stop_token)
-            mm = out.multimodal_output
-            if mm is None:
-                continue
-            audio = _extract_audio(mm)
-            if audio is None:
-                continue
-            sr_t = mm.get("sr")
-            if hasattr(sr_t, "item"):
-                assert int(sr_t.item()) == SAMPLE_RATE
-            waveform = audio.cpu().float()
-        assert waveform is not None, "no audio output produced"
-        by_index[index] = _Clip(wav=waveform, committed_frames=frames)
-
-    assert sorted(by_index) == list(range(len(texts))), (
-        f"request ids did not cover every submitted prompt exactly once: got {sorted(by_index)}"
-    )
-    return [by_index[i] for i in range(len(texts))]
-
-
-def _synthesize(omni: Omni, text: str) -> _Clip:
-    return _synthesize_all(omni, [text])[0]
-
-
-def _assert_clip_is_sane(clip: _Clip) -> None:
-    """The checks every clip has to pass, whatever produced it."""
-    wav = clip.wav
-    assert wav.dim() == 1, f"expected mono 1-D, got shape {tuple(wav.shape)}"
-    assert wav.numel() > 0, "empty waveform"
-    assert torch.isfinite(wav).all(), "non-finite samples"
-
-    # Frame conservation against the engine's own token count, not just
-    # divisibility. This is what catches a payload carrying non-audio data or a
-    # dropped tail; both pass every other check here.
-    assert wav.numel() == clip.committed_frames * SAMPLES_PER_FRAME, (
-        f"{wav.numel()} samples for {clip.committed_frames} committed frames — "
-        f"expected exactly {clip.committed_frames * SAMPLES_PER_FRAME}"
-    )
-
-    # Anything far outside the waveform range is not codec output.
-    assert float(wav.abs().max()) <= 1.5, f"peak {float(wav.abs().max()):.2f} is out of range"
-
-
-def _assert_stopped_on_its_own(clip: _Clip) -> None:
-    """The request must stop on its own rather than run to ``max_tokens``.
-
-    Separate from the structural checks because it reads the stop head's
-    judgement, not the payload's shape: under ``load_format: dummy`` that head
-    is random and never fires, so every clip runs to the token cap.
-    """
-    seconds = clip.wav.numel() / SAMPLE_RATE
-    assert 0.5 < seconds < 30.0, f"implausible duration {seconds:.2f}s — did STOP fire?"
-
-
-def _transcribe(wav: torch.Tensor, tmp_dir: str, name: str, model_size: str = _ASR_MODEL) -> str:
-    path = str(Path(tmp_dir) / f"{name}.wav")
-    sf.write(path, wav.numpy(), SAMPLE_RATE)
-    return convert_audio_file_to_text(path, model_size=model_size, language="en").lower()
-
-
-def _transcribe_for(wav: torch.Tensor, tmp_dir: str, name: str, accepts) -> str:
-    """Transcribe with the fast model, retrying once with the stronger one.
-
-    ``accepts`` decides whether the fast transcript is usable. Escalating only
-    on a miss keeps the average clip at ``_ASR_MODEL`` while a genuine model
-    artifact still fails the caller's assertion, since the stronger ASR
-    mistranscribes it too.
-    """
-    transcript = _transcribe(wav, tmp_dir, name)
-    if accepts(transcript):
-        return transcript
-    return _transcribe(wav, tmp_dir, name, model_size=_ASR_ESCALATION_MODEL)
-
-
-@hardware_test(res={"cuda": "L4"}, num_cards=1)
-@pytest.mark.parametrize(
-    ("text", "keyword"),
-    [
-        # The canonical path: one request, one transcript, the cheapest check
-        # that reads the audio as speech.
-        pytest.param(
-            "Hello, this is Gepard speaking.",
-            "hello",
-            marks=pytest.mark.advanced_model,
-            id="canonical",
-        ),
-        # No keyword: this row carries the structural and stopping checks over a
-        # second text and leaves the transcript to the row above.
-        pytest.param(
-            "The quick brown fox jumps over the lazy dog.",
-            None,
-            marks=pytest.mark.full_model,
-            id="second_text",
-        ),
-    ],
-)
-def test_gepard_offline_zero_shot(omni_runner, run_level: str, text: str, keyword: str | None) -> None:
-    """Zero-shot synthesis produces a finite mono waveform at 22.05 kHz.
-
-    The transcript check is the one assertion here that reads the audio as
-    speech rather than as a tensor. It stands in for a parity comparison, which
-    needs a reference implementation this PR cannot run — noise, a repeated
-    loop or the wrong clip all pass every numeric check above it.
-    """
-    clip = _synthesize(omni_runner, text)
-    _assert_clip_is_sane(clip)
-
-    if run_level not in {"advanced_model", "full_model"}:
-        return  # dummy weights: the audio is structural only, so do not read it
-
-    _assert_stopped_on_its_own(clip)
-    if keyword is None:
-        return
-
-    with tempfile.TemporaryDirectory() as tmp_dir:
-        transcript = _transcribe_for(clip.wav, tmp_dir, "zero_shot", lambda t: keyword in t)
-    assert keyword in transcript, f"expected {keyword!r} in transcript, got {transcript!r}"
-
-
-@pytest.mark.full_model
-@hardware_test(res={"cuda": "L4"}, num_cards=1)
-def test_gepard_offline_long_text_skips_repetition(omni_runner, run_level: str) -> None:
-    """The other branch of the prompt layout, end to end.
-
-    A text at or above ``apply_below`` tokens is voiced once; shorter ones are
-    repeated to carry enough mass against the speaker prefix. The layout
-    assertion keeps this row honest — if a tokenizer change pushed this text
-    back under the threshold it would silently retest the repeated path.
-    """
-    prompt_ids = _build_request(_LONG_TEXT)["prompt_token_ids"]
-    assert prompt_ids.count(_config().start_of_text) == 1, "_LONG_TEXT is no longer above apply_below"
-
-    clip = _synthesize(omni_runner, _LONG_TEXT)
-    _assert_clip_is_sane(clip)
-
-    if run_level not in {"advanced_model", "full_model"}:
-        return
-
-    _assert_stopped_on_its_own(clip)
-    with tempfile.TemporaryDirectory() as tmp_dir:
-        transcript = _transcribe_for(clip.wav, tmp_dir, "long_text", lambda t: "river" in t)
-    assert "river" in transcript, f"expected 'river' in transcript, got {transcript!r}"
-
-
-@pytest.mark.slow
-@hardware_test(res={"cuda": "L4"}, num_cards=1)
-def test_gepard_offline_concurrent_requests_stay_isolated(omni_runner, run_level: str) -> None:
-    """Four requests in one batch, each holding its own generation state.
-
-    The prompts are deliberately unlike each other: cross-talk between the
-    per-request frame history, the codec window or the output routing shows up
-    as a transcript matching the wrong prompt, which similar prompts would
-    hide. ``max_num_seqs`` is 4, so these really do run together.
-    """
-    texts = list(_DISTINGUISHABLE_PROMPTS)
-    clips = _synthesize_all(omni_runner, texts)
-
-    for clip in clips:
-        _assert_clip_is_sane(clip)
-
-    if run_level not in {"advanced_model", "full_model"}:
-        return  # dummy weights: the audio is structural only, so do not read it
-
-    for clip in clips:
-        _assert_stopped_on_its_own(clip)
-
-    # All four clips, not just the transcribed ones: delivering one request's
-    # audio for another leaves two byte-identical waveforms.
-    for i in range(len(clips)):
-        for j in range(i + 1, len(clips)):
-            assert not torch.equal(clips[i].wav, clips[j].wav), (
-                f"clips {i} and {j} are the same waveform — one request's audio was delivered for another"
-            )
-
-    # Each clip must say what its own prompt asked for: accepting any of the
-    # four would pass a stable permutation, which is what a mix-up produces.
-    keywords = set(_DISTINGUISHABLE_PROMPTS.values())
-    with tempfile.TemporaryDirectory() as tmp_dir:
-        for i, clip in enumerate(clips[:_TRANSCRIBED_CLIPS]):
-            expected = _DISTINGUISHABLE_PROMPTS[texts[i]]
-            # Escalate on the expected word being absent, not on "no keyword
-            # at all" -- a neighbour's word is the failure under test.
-            transcript = _transcribe_for(clip.wav, tmp_dir, f"concurrent_{i}", lambda t: expected in t)
-            present = sorted(k for k in keywords if k in transcript)
-            assert present == [expected], (
-                f"clip for {texts[i]!r} transcribed as {transcript!r}: expected {expected!r} and "
-                f"no other prompt's keyword, found {present} — requests bled into each other"
-            )
@@ -39,7 +39,8 @@ def _extract_audio(multimodal_output: dict) -> torch.Tensor:
    return audio


-@pytest.mark.slow
+@pytest.mark.core_model
+@pytest.mark.advanced_model
@pytest.mark.tts
@hardware_test(res={"cuda": "L4"}, num_cards=1)
 def test_voxcpm2_zero_shot_001(omni_runner: OmniRunner) -> None:
@@ -52,7 +53,8 @@ def test_voxcpm2_zero_shot_001(omni_runner: OmniRunner) -> None:
    assert 0.5 < duration_s < 30.0, f"Audio duration out of range: {duration_s:.2f}s"


-@pytest.mark.slow
+@pytest.mark.core_model
+@pytest.mark.advanced_model
@pytest.mark.tts
@hardware_test(res={"cuda": "L4"}, num_cards=1)
 def test_voxcpm2_voice_clone_002(omni_runner: OmniRunner) -> None:
@@ -93,7 +95,8 @@ def test_voxcpm2_voice_clone_002(omni_runner: OmniRunner) -> None:
    assert 0.5 < duration_s < 30.0, f"Audio duration out of range: {duration_s:.2f}s"


-@pytest.mark.slow
+@pytest.mark.core_model
+@pytest.mark.advanced_model
@pytest.mark.tts
@hardware_test(res={"cuda": "L4"}, num_cards=1)
 def test_voxcpm2_prefill_decode_mixed_batch_003(omni_runner: OmniRunner) -> None:
@@ -11,7 +11,7 @@ MODEL = "Tongyi-MAI/Z-Image-Turbo"
 PROMPT = "A high-detail studio photo of an orange tabby cat sitting on a laptop keyboard."


-@pytest.mark.slow
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@hardware_test(res={"cuda": "L4", "rocm": "MI325", "xpu": "B60"}, num_cards={"cuda": 1, "rocm": 1, "xpu": 2})
@pytest.mark.parametrize("omni_runner", [(MODEL, None)], indirect=True)
@@ -75,7 +75,8 @@ def _build_img2img_messages(prompt: str, image_b64: str) -> list[dict]:
    ]


-@pytest.mark.slow
+@pytest.mark.core_model
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@hardware_test(res={"cuda": "H100"})
@pytest.mark.parametrize("omni_server", test_params, indirect=True)
@@ -97,7 +98,8 @@ def test_bagel_text2img_online(omni_server, openai_client) -> None:
    openai_client.send_diffusion_request(request_config)


-@pytest.mark.slow
+@pytest.mark.core_model
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@hardware_test(res={"cuda": "H100", "rocm": "MI325"})
@pytest.mark.parametrize("omni_server", test_params, indirect=True)
@@ -59,7 +59,8 @@ _REF_TEXT = "Okay. Yeah. I resent you. I love you. I respect you. But you know w
 class TestHiggsAudioV3OnlineHappyPath:
    """Plain-text -> audio happy paths against the live HTTP server."""

-    @pytest.mark.slow
+    @pytest.mark.core_model
+    @pytest.mark.advanced_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_plain_text_wav(self, omni_server, openai_client) -> None:
@@ -78,7 +79,7 @@ class TestHiggsAudioV3OnlineHappyPath:
            }
        )

-    @pytest.mark.slow
+    @pytest.mark.core_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_plain_text_with_max_new_tokens(self, omni_server, openai_client) -> None:
@@ -96,7 +97,8 @@ class TestHiggsAudioV3OnlineHappyPath:
        )

    @pytest.mark.skip(reason="issue#4411")
-    @pytest.mark.slow
+    @pytest.mark.core_model
+    @pytest.mark.advanced_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_concurrent_plain_text(self, omni_server, openai_client) -> None:
@@ -113,7 +115,7 @@ class TestHiggsAudioV3OnlineHappyPath:
        )

    @pytest.mark.skip(reason="issue#4411")
-    @pytest.mark.slow
+    @pytest.mark.advanced_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_plain_text_pcm_streaming(self, omni_server, openai_client) -> None:
@@ -145,7 +147,7 @@ class TestHiggsAudioV3OnlineHappyPath:
        )

    @pytest.mark.skip(reason="issue#4411")
-    @pytest.mark.slow
+    @pytest.mark.advanced_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_concurrent_pcm_streaming(self, omni_server, openai_client) -> None:
@@ -187,7 +189,7 @@ class TestHiggsAudioV3OnlineInlineControlTokens:
    They do not assert audio quality (out of scope for CI).
    """

-    @pytest.mark.slow
+    @pytest.mark.core_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_inline_emotion_and_expressive(self, omni_server, openai_client) -> None:
@@ -203,7 +205,7 @@ class TestHiggsAudioV3OnlineInlineControlTokens:
            }
        )

-    @pytest.mark.slow
+    @pytest.mark.core_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_inline_style_whispering(self, omni_server, openai_client) -> None:
@@ -219,7 +221,7 @@ class TestHiggsAudioV3OnlineInlineControlTokens:
            }
        )

-    @pytest.mark.slow
+    @pytest.mark.core_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_inline_prosody_speed_and_pitch(self, omni_server, openai_client) -> None:
@@ -237,7 +239,7 @@ class TestHiggsAudioV3OnlineInlineControlTokens:
            }
        )

-    @pytest.mark.slow
+    @pytest.mark.core_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_inline_pause_mid_text(self, omni_server, openai_client) -> None:
@@ -253,7 +255,7 @@ class TestHiggsAudioV3OnlineInlineControlTokens:
            }
        )

-    @pytest.mark.slow
+    @pytest.mark.core_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_inline_sfx_with_onomatopoeia(self, omni_server, openai_client) -> None:
@@ -277,7 +279,8 @@ class TestHiggsAudioV3OnlineInlineControlTokens:
 class TestHiggsAudioV3OnlineVoiceClone:
    """Voice clone via the two payload shapes the model serves."""

-    @pytest.mark.slow
+    @pytest.mark.core_model
+    @pytest.mark.advanced_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_voice_clone_ref_audio_ref_text(self, omni_server, openai_client) -> None:
@@ -294,7 +297,7 @@ class TestHiggsAudioV3OnlineVoiceClone:
            }
        )

-    @pytest.mark.slow
+    @pytest.mark.core_model
    @pytest.mark.tts
    @hardware_test(res={"cuda": "L4"}, num_cards=1)
    def test_voice_clone_references_alias(self, omni_server, openai_client) -> None:
@@ -1,89 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-
-"""Online serving smokes for all four canonical LTX-2.5 pipelines."""
-
-import os
-
-import pytest
-
-from tests.helpers import skip_if_gated_repo_inaccessible
-from tests.helpers.mark import hardware_marks
-from tests.helpers.media import generate_synthetic_image
-from tests.helpers.runtime import OmniServer, OmniServerParams, OpenAIClientHandler
-
-os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"
-
-DEFAULT_MODEL = "Lightricks/LTX-2.5-Diffusers"
-DEFAULT_REVISION = "a6de4b5354f078db24d9cf4778c14846788aea3d"
-MODEL = os.getenv("VLLM_TEST_LTX25_MODEL", DEFAULT_MODEL)
-MODEL_REVISION = os.getenv("VLLM_TEST_LTX25_MODEL_REVISION", DEFAULT_REVISION if MODEL == DEFAULT_MODEL else "")
-PROMPT = "A red fox walks through a snowy forest while the camera remains fixed."
-
-pytestmark = [pytest.mark.diffusion, pytest.mark.full_model]
-SINGLE_CARD_MARKS = hardware_marks(res={"cuda": "H100"})
-
-
-@pytest.fixture(scope="module", autouse=True)
-def require_ltx25_model_access() -> None:
-    if not os.path.isdir(MODEL):
-        skip_if_gated_repo_inaccessible(
-            MODEL,
-            revision=MODEL_REVISION or None,
-            filename="model_index.json",
-        )
-
-
-def _server(model_class_name: str) -> OmniServerParams:
-    return OmniServerParams(
-        model=MODEL,
-        server_args=[
-            *(["--revision", MODEL_REVISION] if MODEL_REVISION else []),
-            "--model-class-name",
-            model_class_name,
-            "--enforce-eager",
-            "--enable-layerwise-offload",
-            "--diffusion-attention-backend",
-            "CUDNN_ATTN",
-        ],
-    )
-
-
-def _cases():
-    return [
-        pytest.param(_server("LTX2Pipeline"), 30, id="full_one_stage", marks=SINGLE_CARD_MARKS),
-        pytest.param(_server("LTX2TwoStagePipeline"), 30, id="full_two_stage", marks=SINGLE_CARD_MARKS),
-        pytest.param(_server("LTX2DistilledOneStagePipeline"), 8, id="distilled_one_stage", marks=SINGLE_CARD_MARKS),
-        pytest.param(_server("LTX2DistilledTwoStagePipeline"), 8, id="distilled_two_stage", marks=SINGLE_CARD_MARKS),
-    ]
-
-
-@pytest.mark.parametrize(("omni_server", "num_inference_steps"), _cases(), indirect=["omni_server"])
-def test_ltx25_pipeline_entries(
-    omni_server: OmniServer,
-    num_inference_steps: int,
-    openai_client: OpenAIClientHandler,
-    subtests,
-) -> None:
-    """Generate T2V and first-frame I2V through the synchronous video API."""
-    for task in ("t2v", "i2v"):
-        with subtests.test(task=task):
-            request_config = {
-                "model": omni_server.model,
-                "form_data": {
-                    "model": omni_server.model,
-                    "prompt": PROMPT,
-                    "height": 128,
-                    "width": 128,
-                    "num_frames": 9,
-                    "fps": 24,
-                    "num_inference_steps": num_inference_steps,
-                    "seed": 42,
-                },
-            }
-            if task == "i2v":
-                request_config["image_reference"] = (
-                    f"data:image/jpeg;base64,{generate_synthetic_image(512, 512)['base64']}"
-                )
-
-            openai_client.send_video_diffusion_request(request_config)
@@ -14,10 +14,9 @@ os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"
 import pytest

 from tests.helpers.mark import hardware_test
-from tests.helpers.media import get_asset_path, load_test_audio_data_url
+from tests.helpers.media import load_test_audio_data_url
 from tests.helpers.runtime import OmniServerParams
 from tests.helpers.stage_config import get_deploy_config_path
-from vllm_omni.entrypoints.openai.serving_speech import _DEFAULT_VOICE_NAME

 try:
    from transformers import HiggsAudioV2TokenizerModel  # noqa: F401
@@ -141,121 +140,3 @@ class TestOmniVoiceVoiceCloning:
            "min_audio_bytes": _DEFAULT_MIN_AUDIO_BYTES,
        }
        openai_client.send_audio_speech_request(request_config)
-
-
-@hardware_test(res={"cuda": "L4"}, num_cards=1)
-@pytest.mark.parametrize("omni_server", TEST_PARAMS, indirect=True)
-@pytest.mark.parametrize("voice", [_DEFAULT_VOICE_NAME, {"id": _DEFAULT_VOICE_NAME}])
-def test_speech_with_voice_param_accepted(omni_server, openai_client, voice) -> None:
-    """The default voice should be accepted as both a plain string and a
-    VoiceID dict, even when the model has no uploaded speakers."""
-    request_config = {
-        "model": omni_server.model,
-        "input": get_prompt("text"),
-        "voice": voice,
-        "response_format": "wav",
-        "timeout": 180.0,
-        "min_audio_bytes": _DEFAULT_MIN_AUDIO_BYTES,
-    }
-    openai_client.send_audio_speech_request(request_config)
-
-
-@hardware_test(res={"cuda": "L4"}, num_cards=1)
-@pytest.mark.parametrize("omni_server", TEST_PARAMS, indirect=True)
-def test_unknown_voice_rejected(omni_server, openai_client) -> None:
-    """An unrecognized voice name should be rejected with a 400."""
-    request_config = {
-        "model": omni_server.model,
-        "input": get_prompt("text"),
-        "voice": "nonexistent_voice",
-        "status_code": 400,
-        "err_message": "Invalid voice",
-    }
-    openai_client.send_audio_speech_request(request_config)
-
-
-@hardware_test(res={"cuda": "L4"}, num_cards=1)
-@pytest.mark.parametrize("omni_server", TEST_PARAMS, indirect=True)
-def test_voice_request_before_and_after_clone_registration(omni_server, openai_client) -> None:
-    """The default voice should always be listed and usable, even after
-    a different cloned voice is registered."""
-    speech_config = {
-        "model": omni_server.model,
-        "input": get_prompt("text"),
-        "voice": _DEFAULT_VOICE_NAME,
-    }
-
-    # 1) "default" should appear in the voices list before any uploads
-    voices = openai_client.send_audio_voices_list_http_request()[0]
-    assert _DEFAULT_VOICE_NAME in voices.json_body["voices"]
-
-    # 2) Speech with the default voice should succeed
-    openai_client.send_audio_speech_request(speech_config)
-
-    # 3) Register a cloned voice under a different name
-    register_config = {
-        "data": {"name": "bar", "consent": "test-consent"},
-        "files": {
-            "audio_sample": (
-                "clone_2.wav",
-                get_asset_path("qwen3_tts/clone_2.wav").read_bytes(),
-                "audio/wav",
-            ),
-        },
-    }
-    openai_client.send_audio_voices_create_http_request(register_config)
-
-    # 4) "default" should still be listed alongside the new voice
-    voices = openai_client.send_audio_voices_list_http_request()[0]
-    assert _DEFAULT_VOICE_NAME in voices.json_body["voices"]
-    assert "bar" in voices.json_body["voices"]
-
-    # 5) Speech with the default voice should still succeed
-    openai_client.send_audio_speech_request(speech_config)
-
-
-@hardware_test(res={"cuda": "L4"}, num_cards=1)
-@pytest.mark.parametrize("omni_server", TEST_PARAMS, indirect=True)
-def test_registered_default_voice_overrides_placeholder(omni_server, openai_client) -> None:
-    """When a real voice is uploaded under the default name, it should be
-    used as a cloned voice; after deletion, the placeholder behavior resumes."""
-    speech_config = {
-        "model": omni_server.model,
-        "input": get_prompt("text"),
-        "voice": _DEFAULT_VOICE_NAME,
-    }
-
-    # 1) Placeholder default works before any registration
-    openai_client.send_audio_speech_request(speech_config)
-
-    # 2) Register a real voice under the default name
-    register_config = {
-        "data": {"name": _DEFAULT_VOICE_NAME, "consent": "test-consent"},
-        "files": {
-            "audio_sample": (
-                "clone_2.wav",
-                get_asset_path("qwen3_tts/clone_2.wav").read_bytes(),
-                "audio/wav",
-            ),
-        },
-    }
-    openai_client.send_audio_voices_create_http_request(register_config)
-
-    # 3) After creating the voice with the default name, it's visible as an uploaded_voice
-    voices = openai_client.send_audio_voices_list_http_request()[0]
-    assert _DEFAULT_VOICE_NAME in voices.json_body["voices"]
-    uploaded_names = [v["name"] for v in voices.json_body["uploaded_voices"]]
-    assert _DEFAULT_VOICE_NAME in uploaded_names
-    openai_client.send_audio_speech_request(speech_config)
-
-    # 4) Delete the uploaded default voice
-    openai_client.send_audio_voices_delete_http_request(
-        {"name": _DEFAULT_VOICE_NAME},
-    )
-
-    # 5) Ensure the placeholder behavior is consistent, but it's not an uploaded voice now
-    voices = openai_client.send_audio_voices_list_http_request()[0]
-    assert _DEFAULT_VOICE_NAME in voices.json_body["voices"]
-    uploaded_names = [v["name"] for v in voices.json_body["uploaded_voices"]]
-    assert _DEFAULT_VOICE_NAME not in uploaded_names
-    openai_client.send_audio_speech_request(speech_config)
@@ -6,7 +6,7 @@ Online serving tests for the Qwen-Image-Edit family (image-to-image via chat com

 - ``test_single_image_to_image_001``: ``Qwen/Qwen-Image-Edit`` — one reference image, fixed 512×512.
 - ``test_multi_images_to_image_001``: ``Qwen/Qwen-Image-Edit-2511`` — two reference images, fixed 512×512.
- ``test_different_sizes_001``: ``Qwen/Qwen-Image-Edit-2511`` only, ``slow`` — mixed input
+- ``test_different_sizes_001``: ``Qwen/Qwen-Image-Edit-2511`` only, ``advanced_model`` — mixed input
  resolutions; ``extra_body`` uses per-output ``width``/``height`` lists; the test client sends one
  scalar-size request per list index in parallel and merges images (see ``OpenAIClientHandler.send_diffusion_request``).

@@ -52,7 +52,7 @@ def _get_diffusion_feature_cases(model: str):
    ]


-@pytest.mark.slow
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@pytest.mark.parametrize(
    "omni_server",
@@ -82,7 +82,8 @@ def test_single_image_to_image_001(omni_server: OmniServer, openai_client: OpenA
    openai_client.send_diffusion_request(request_config)


-@pytest.mark.slow
+@pytest.mark.core_model
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@pytest.mark.parametrize(
    "omni_server",
@@ -113,7 +114,7 @@ def test_multi_images_to_image_001(omni_server: OmniServer, openai_client: OpenA
    openai_client.send_diffusion_request(request_config)


-@pytest.mark.slow
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@pytest.mark.parametrize(
    "omni_server",
@@ -44,7 +44,8 @@ def _get_diffusion_feature_cases(model: str):
    ]


-@pytest.mark.slow
+@pytest.mark.core_model
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@pytest.mark.parametrize("omni_server", _get_diffusion_feature_cases(MODEL), indirect=True)
 def test_single_image_to_image_001(omni_server: OmniServer, openai_client: OpenAIClientHandler):
@@ -46,7 +46,8 @@ tts_server_params = [
 ]


-@pytest.mark.slow
+@pytest.mark.advanced_model
+@pytest.mark.core_model
@pytest.mark.tts
@hardware_test(res={"cuda": "L4"}, num_cards=1)
@pytest.mark.parametrize("omni_server", tts_server_params, indirect=True)
@@ -71,7 +72,7 @@ def test_text_to_audio_001(omni_server, openai_client) -> None:
    openai_client.send_audio_speech_request(request_config, request_num=MAX_CONCURRENT)


-@pytest.mark.slow
+@pytest.mark.advanced_model
@pytest.mark.tts
@hardware_test(res={"cuda": "L4"}, num_cards=1)
@pytest.mark.parametrize("omni_server", tts_server_params, indirect=True)
@@ -32,7 +32,7 @@ def _get_diffusion_feature_cases():
    ]


-@pytest.mark.slow
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@pytest.mark.parametrize(
    "omni_server",
@@ -60,7 +60,7 @@ def test_basic_001(
    openai_client.send_diffusion_request(request_config)


-@pytest.mark.slow
+@pytest.mark.advanced_model
@pytest.mark.diffusion
@pytest.mark.parametrize(
    "omni_server",
@@ -161,17 +161,6 @@ def _make_diffusion_plan(
    )


-def _make_stage_runtime() -> StageRuntime:
-    return StageRuntime(
-        stage_configs=[],
-        model="dummy-model",
-        config_path="dummy-config",
-        stage_init_timeout=1,
-        diffusion_batch_size=1,
-        async_chunk=False,
-    )
-
-
 def test_stage_engine_core_client_module_reload_keeps_forward_refs_deferred():
    """Regression test for forward references in make_async_mp_client."""
    import vllm_omni.engine.stage_engine_core_client as client_mod
@@ -244,48 +233,6 @@ def test_collect_initialized_clients_for_cleanup_deduplicates_clients():
    assert cleanup_clients == [shared, extra]


-def test_initialize_local_diffusion_replica_scopes_runtime_env(monkeypatch):
-    import vllm_omni.engine.stage_runtime as runtime_mod
-    from vllm_omni.platforms import current_omni_platform
-
-    runtime = _make_stage_runtime()
-    plan = _make_diffusion_plan(0, stage_id=0).replicas[0]
-
-    runtime_env_var = "VLLM_OMNI_TEST_STAGE_RUNTIME_ENV"
-    device_env_var = current_omni_platform.device_control_env_var
-    runtime._init_visible_devices_baseline = "0,1"
-    plan.metadata.runtime_cfg = {
-        "devices": "0",
-        "env": {runtime_env_var: "stage-value"},
-    }
-    monkeypatch.delenv(runtime_env_var, raising=False)
-    monkeypatch.setenv(device_env_var, "0,1")
-
-    captured: dict[str, str | None] = {}
-    monkeypatch.setattr(runtime_mod, "inject_kv_stage_info", lambda *_: None)
-
-    def _capture_launch_diffusion_stage_replica(**_kwargs):
-        captured["runtime_env"] = os.environ.get(runtime_env_var)
-        captured["device_env"] = os.environ.get(device_env_var)
-        raise RuntimeError("stop after capturing launch environment")
-
-    monkeypatch.setattr(
-        runtime_mod,
-        "launch_diffusion_stage_replica",
-        _capture_launch_diffusion_stage_replica,
-    )
-
-    with pytest.raises(RuntimeError, match="stop after capturing launch environment"):
-        runtime._initialize_local_diffusion_replica(plan, stage_init_timeout=1)
-
-    assert captured == {
-        "runtime_env": "stage-value",
-        "device_env": "0",
-    }
-    assert runtime_env_var not in os.environ
-    assert os.environ[device_env_var] == "0,1"
-
-
 def test_initialize_local_diffusion_replica_restores_device_visibility_after_local_init(monkeypatch):
    import vllm_omni.engine.stage_runtime as runtime_mod
    from vllm_omni.engine.stage_engine_startup import StageReplicaResources
@@ -364,136 +311,6 @@ def test_initialize_local_diffusion_replica_passes_stage_init_timeout_and_inline
    }


-def test_initialize_local_llm_replica_scopes_runtime_env(monkeypatch):
-    import vllm_omni.engine.stage_runtime as runtime_mod
-
-    runtime = _make_stage_runtime()
-    plan = _make_llm_plan(0, stage_id=0, vllm_config=object()).replicas[0]
-    plan.engine_args_dict = {}
-
-    runtime_env_var = "VLLM_OMNI_TEST_STAGE_RUNTIME_ENV"
-    plan.metadata.runtime_cfg = {
-        "devices": "0",
-        "env": {runtime_env_var: "stage-value"},
-    }
-    monkeypatch.setenv(runtime_env_var, "parent-value")
-
-    captured: dict[str, str | None] = {}
-
-    @contextlib.contextmanager
-    def _capture_launch_stage_replica(*, stage_visible_devices, **_kwargs):
-        captured["runtime_env"] = os.environ.get(runtime_env_var)
-        captured["stage_visible_devices"] = stage_visible_devices
-        yield None
-
-    monkeypatch.setattr(runtime_mod, "acquire_device_locks", lambda *_args, **_kwargs: [])
-    monkeypatch.setattr(runtime_mod, "launch_stage_replica", _capture_launch_stage_replica)
-
-    with pytest.raises(RuntimeError, match="launcher returned no resources"):
-        runtime._initialize_local_llm_replica(plan, stage_init_timeout=1)
-
-    assert captured == {
-        "runtime_env": "stage-value",
-        "stage_visible_devices": "0",
-    }
-    assert os.environ[runtime_env_var] == "parent-value"
-
-
-def test_initialize_diffusion_stage_applies_client_batch_size_to_engine(monkeypatch):
-    import vllm_omni.diffusion.stage_diffusion_client as client_mod
-    import vllm_omni.engine.stage_init_utils as init_mod
-
-    od_config = types.SimpleNamespace(max_num_seqs=1)
-    captured: dict[str, object] = {}
-    monkeypatch.setattr(init_mod, "build_diffusion_config", lambda *args: od_config)
-
-    def _capture_client(model, config, metadata, stage_init_timeout, batch_size, use_inline):
-        captured.update(
-            model=model,
-            config=config,
-            metadata=metadata,
-            stage_init_timeout=stage_init_timeout,
-            batch_size=batch_size,
-            use_inline=use_inline,
-        )
-        return object()
-
-    monkeypatch.setattr(client_mod, "create_diffusion_client", _capture_client)
-    metadata = _make_diffusion_metadata(0)
-
-    init_mod.initialize_diffusion_stage(
-        0,
-        "dummy-model",
-        types.SimpleNamespace(),
-        metadata,
-        stage_init_timeout=12,
-        batch_size=4,
-        use_inline=True,
-    )
-
-    assert od_config.max_num_seqs == 4
-    assert captured == {
-        "model": "dummy-model",
-        "config": od_config,
-        "metadata": metadata,
-        "stage_init_timeout": 12,
-        "batch_size": 4,
-        "use_inline": True,
-    }
-
-
-def test_launch_diffusion_stage_replica_applies_batch_size_to_config(monkeypatch):
-    import vllm_omni.diffusion.stage_diffusion_client as client_mod
-    import vllm_omni.diffusion.stage_diffusion_proc as proc_mod
-    import vllm_omni.engine.stage_engine_startup as startup_mod
-
-    od_config = types.SimpleNamespace(max_num_seqs=1, parallel_config=types.SimpleNamespace(world_size=1))
-    monkeypatch.setattr(startup_mod, "build_diffusion_config", lambda *args: od_config)
-    monkeypatch.setattr(startup_mod, "acquire_device_locks", lambda *args: [])
-    monkeypatch.setattr(
-        startup_mod,
-        "register_stage_with_omni_master",
-        lambda **kwargs: types.SimpleNamespace(
-            handshake_address="tcp://127.0.0.1:26001",
-            input_address="tcp://127.0.0.1:26002",
-            output_address="tcp://127.0.0.1:26003",
-        ),
-    )
-
-    omni_master_server = types.SimpleNamespace(
-        address="127.0.0.1",
-        port=25000,
-        release_route_port_reservations=lambda *args, **kwargs: None,
-    )
-    proc_manager = types.SimpleNamespace(
-        addresses=types.SimpleNamespace(
-            inputs=["tcp://127.0.0.1:26002"],
-            outputs=["tcp://127.0.0.1:26003"],
-        )
-    )
-    monkeypatch.setattr(proc_mod, "StageDiffusionProcManager", lambda **kwargs: proc_manager)
-    sentinel_client = object()
-    monkeypatch.setattr(
-        client_mod.StageDiffusionClient,
-        "from_addresses",
-        lambda metadata, **kwargs: sentinel_client,
-    )
-
-    result, resources = startup_mod.launch_diffusion_stage_replica(
-        model="dummy-model",
-        stage_config=types.SimpleNamespace(),
-        metadata=types.SimpleNamespace(stage_id=0),
-        stage_init_timeout=12,
-        batch_size=4,
-        use_inline=False,
-        omni_master_server=omni_master_server,
-    )
-
-    assert result is sentinel_client
-    assert od_config.max_num_seqs == 4
-    assert resources.manager is proc_manager
-
-
 def test_stage_runtime_initializes_stage_pools(monkeypatch):
    import vllm_omni.engine.stage_runtime as runtime_mod

@@ -270,56 +270,3 @@ class TestMsgpackRoundTrip:
        # Content merged from legacy nested key
        assert result.prompt == "legacy prompt"
        assert result.prompt_token_ids == [1, 2]
-
-
-class TestRequestOutputFieldParity:
-    """``OmniRequestOutput`` must expose everything ``RequestOutput`` sets.
-
-    ``RequestOutput`` is a plain class, so the dataclass-generated ``__init__``
-    replaces its ``__init__`` and the inherited attributes only exist because
-    they are redeclared as fields. Any vLLM field that is not redeclared simply
-    does not exist on an omni output, and vLLM's own serving code reads those
-    attributes unconditionally — ``/v1/completions`` returned HTTP 500 with
-    ``AttributeError: 'OmniRequestOutput' object has no attribute
-    'ec_transfer_params'`` for exactly that reason. This test fails on the next
-    vLLM bump that adds a field, instead of leaving it for an endpoint to hit.
-    """
-
-    def test_carries_every_attribute_request_output_sets(self) -> None:
-        reference = _make_text_request_output()
-        omni = OmniRequestOutput(request_id="req-parity")
-
-        missing = sorted(name for name in vars(reference) if not name.startswith("_") and not hasattr(omni, name))
-        assert not missing, (
-            f"OmniRequestOutput is missing RequestOutput attributes {missing}; "
-            "redeclare them as dataclass fields in vllm_omni/outputs/__init__.py"
-        )
-
-    def test_transfer_param_fields_default_to_none(self) -> None:
-        omni = OmniRequestOutput(request_id="req-defaults")
-        assert omni.kv_transfer_params is None
-        assert omni.ec_transfer_params is None
-        assert omni.num_cached_tokens is None
-        assert omni.num_cache_creation_tokens is None
-
-    def test_transfer_param_fields_are_copied_from_the_stage_output(self) -> None:
-        """Declaring the fields is not enough — they must also be copied.
-
-        ``from_stage_output`` only carries the names listed in
-        ``_REQUEST_OUTPUT_CONTENT_ATTRS``. With the fields declared but absent
-        from that list the endpoint stops raising ``AttributeError`` while
-        silently dropping external-connector metadata and cache-creation
-        accounting, which is harder to notice than the crash it replaced.
-        """
-        source = _make_text_request_output()
-        source.num_cache_creation_tokens = 7
-        source.ec_transfer_params = {"connector": "shm", "handle": "abc"}
-        source.num_cached_tokens = 3
-        source.kv_transfer_params = {"role": "producer"}
-
-        omni = OmniRequestOutput.from_stage_output(source, stage_id=1)
-
-        assert omni.num_cache_creation_tokens == 7
-        assert omni.ec_transfer_params == {"connector": "shm", "handle": "abc"}
-        assert omni.num_cached_tokens == 3
-        assert omni.kv_transfer_params == {"role": "producer"}
@@ -1138,8 +1138,7 @@ class TestSingleStageReplicaInitialization:
        runtime._init_visible_devices_baseline = "0"

        mocker.patch.object(runtime_mod, "inject_kv_stage_info")
-        od_config = SimpleNamespace(max_num_seqs=None, parallel_config=SimpleNamespace(world_size=1))
-        mocker.patch("vllm_omni.engine.stage_engine_startup.build_diffusion_config", return_value=od_config)
+        mocker.patch("vllm_omni.engine.stage_engine_startup.build_diffusion_config", return_value="diffusion-config")
        mock_register = mocker.patch(
            "vllm_omni.engine.stage_engine_startup.register_stage_with_omni_master",
            return_value=StageRegistrationResponse(
@@ -1175,7 +1174,6 @@ class TestSingleStageReplicaInitialization:
                os.environ[device_env_var] = prev_device_env

        assert result is sentinel_client
-        assert od_config.max_num_seqs == 4
        mock_register.assert_called_once_with(
            omni_master_address="127.0.0.1",
            omni_master_port=25000,
@@ -1185,7 +1183,7 @@ class TestSingleStageReplicaInitialization:
        )
        mock_manager.assert_called_once_with(
            model="fake-model",
-            od_config=od_config,
+            od_config="diffusion-config",
            stage_init_timeout=60,
            handshake_address="tcp://127.0.0.1:26001",
            addresses=mocker.ANY,
@@ -1232,8 +1230,7 @@ class TestSingleStageReplicaInitialization:
        runtime._init_visible_devices_baseline = "0"

        mocker.patch.object(runtime_mod, "inject_kv_stage_info")
-        od_config = SimpleNamespace(max_num_seqs=None, parallel_config=SimpleNamespace(world_size=1))
-        mocker.patch("vllm_omni.engine.stage_engine_startup.build_diffusion_config", return_value=od_config)
+        mocker.patch("vllm_omni.engine.stage_engine_startup.build_diffusion_config", return_value="diffusion-config")
        mocker.patch(
            "vllm_omni.engine.stage_engine_startup.register_stage_with_omni_master",
            return_value=StageRegistrationResponse(
@@ -32,7 +32,7 @@ from vllm_omni.entrypoints.openai.protocol.videos import (
    VideoParams,
    VideoResponse,
 )
-from vllm_omni.entrypoints.openai.serving_video import OmniOpenAIServingVideo, ReferenceImage
+from vllm_omni.entrypoints.openai.serving_video import OmniOpenAIServingVideo
 from vllm_omni.entrypoints.openai.storage import LocalStorageManager
 from vllm_omni.entrypoints.openai.stores import AsyncDictStore, TaskRegistry
 from vllm_omni.errors import GuardrailViolationError
@@ -461,51 +461,6 @@ def test_i2v_video_generation_resizes_input_to_requested_dimensions(test_client,
    assert input_image.size == (96, 64)


-def test_i2v_resize_policy_can_defer_to_pipeline(monkeypatch):
-    engine = FakeAsyncOmni()
-    engine.get_diffusion_od_config = lambda: SimpleNamespace(
-        model="org/model",
-        model_class_name="ExamplePipeline",
-        revision="pinned-revision",
-    )
-    captured = {}
-
-    def fake_policy(model_class_name, *, model, revision=None):
-        captured.update(
-            model_class_name=model_class_name,
-            model=model,
-            revision=revision,
-        )
-        return True
-
-    monkeypatch.setattr(
-        "vllm_omni.entrypoints.openai.serving_video.should_preserve_reference_image_size",
-        fake_policy,
-    )
-    handler = OmniOpenAIServingVideo.for_diffusion(
-        diffusion_engine=engine,
-        model_name="fallback/model",
-    )
-    image = Image.new("RGB", (48, 32))
-
-    asyncio.run(
-        handler._run_and_extract(
-            VideoGenerationRequest(prompt="A bear playing with yarn.", width=96, height=64),
-            "pipeline-owned-resize",
-            reference_image=ReferenceImage(image),
-        )
-    )
-
-    input_image = engine.captured_prompt["multi_modal_data"]["image"]
-    assert isinstance(input_image, Image.Image)
-    assert input_image.size == (48, 32)
-    assert captured == {
-        "model_class_name": "ExamplePipeline",
-        "model": "org/model",
-        "revision": "pinned-revision",
-    }
-
-
 def test_i2v_extra_params_dimensions_preserve_input_image_geometry(test_client, mocker: MockerFixture):
    image_bytes = _make_test_image_bytes((48, 48))
    mocker.patch(
@@ -10,10 +10,8 @@ ModelScope path.

 import sys
 import types
-from pathlib import Path

 import pytest
-from pytest_mock import MockerFixture
 from vllm import envs

 from vllm_omni.entrypoints import omni_base
@@ -23,64 +21,6 @@ pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
 MODEL_ID = "Qwen/Qwen3-Omni-30B-A3B-Instruct"


-@pytest.mark.parametrize(
-    "model_uri",
-    [
-        "s3://bucket/model",
-        "gs://bucket/model",
-        "az://bucket/model",
-    ],
-)
-def test_omni_snapshot_download_preserves_object_storage_uri(
-    model_uri: str,
-    mocker: MockerFixture,
-    monkeypatch: pytest.MonkeyPatch,
-) -> None:
-    monkeypatch.delenv("VLLM_USE_MODELSCOPE", raising=False)
-    hf_download = mocker.patch.object(omni_base, "download_weights_from_hf_specific")
-
-    assert omni_base.omni_snapshot_download(model_uri) == model_uri
-    hf_download.assert_not_called()
-
-
-def test_omni_snapshot_download_preserves_existing_local_path(
-    tmp_path: Path,
-    mocker: MockerFixture,
-    monkeypatch: pytest.MonkeyPatch,
-) -> None:
-    monkeypatch.delenv("VLLM_USE_MODELSCOPE", raising=False)
-    absolute_model_path = tmp_path / "absolute-model"
-    absolute_model_path.mkdir()
-    relative_model_path = "relative-model"
-    (tmp_path / relative_model_path).mkdir()
-    monkeypatch.chdir(tmp_path)
-    hf_download = mocker.patch.object(omni_base, "download_weights_from_hf_specific")
-
-    assert omni_base.omni_snapshot_download(str(absolute_model_path)) == str(absolute_model_path)
-    assert omni_base.omni_snapshot_download(relative_model_path) == relative_model_path
-    hf_download.assert_not_called()
-
-
-def test_omni_snapshot_download_uses_hf_for_relative_repo_id(
-    mocker: MockerFixture,
-    monkeypatch: pytest.MonkeyPatch,
-) -> None:
-    monkeypatch.delenv("VLLM_USE_MODELSCOPE", raising=False)
-    model_id = "org/model"
-    # Stub the modular-index probe so the test never reaches huggingface.co.
-    hf_file_probe = mocker.patch.object(omni_base, "file_or_path_exists", return_value=False)
-    hf_download = mocker.patch.object(omni_base, "download_weights_from_hf_specific")
-
-    assert omni_base.omni_snapshot_download(model_id) == model_id
-    hf_file_probe.assert_called_once_with(model_id, "modular_model_index.json", revision=None)
-    hf_download.assert_called_once_with(
-        model_name_or_path=model_id,
-        cache_dir=None,
-        allow_patterns=["*"],
-        require_all=True,
-    )
-
-
@pytest.fixture
 def download_backend(monkeypatch: pytest.MonkeyPatch):
    """Run ``omni_snapshot_download`` and report which backend it selected.
@@ -49,9 +49,7 @@ def test_dreamzero_config_sets_model_class_and_policy_config(monkeypatch):
 def test_dreamzero_enrich_config_preserves_explicit_model_class_name(monkeypatch):
    monkeypatch.setattr(
        "vllm.transformers_utils.config.get_hf_file_to_dict",
-        lambda path, _model, **_kwargs: (
-            None if path == "model_index.json" else {"model_type": "vla", "architectures": ["VLA"]}
-        ),
+        lambda path, _model: None if path == "model_index.json" else {"model_type": "vla", "architectures": ["VLA"]},
    )

    od_config = OmniDiffusionConfig(
@@ -22,7 +22,6 @@ from vllm_omni.engine.stage_init_utils import (
 from vllm_omni.entrypoints.utils import (
    _convert_dataclasses_to_dict,
    _filter_dict_like_object,
-    _try_resolve_omni_model_type,
    coerce_param_message_types,
    filter_dataclass_kwargs,
    filter_stages,
@@ -372,48 +371,6 @@ class TestResolveModelConfigPath:
        assert sender["extra"]["connector_get_max_wait_first_chunk"] == 3000
        assert sender["extra"]["connector_get_max_wait"] == 300

-    def test_object_storage_uri_resolves_via_materialized_configs(
-        self,
-        mocker: MockerFixture,
-        tmp_path,
-    ):
-        """An object-storage URI must not reach HF helpers; resolution reads the
-        locally materialized config files instead."""
-        hf_config = SimpleNamespace(
-            model_type="qwen3_omni_moe",
-            architectures=["Qwen3OmniMoeForConditionalGeneration"],
-        )
-        uri = "s3://qwen3-tts-models/Qwen3-Omni-30B-A3B-Instruct"
-        materialize = mocker.patch(
-            "vllm_omni.entrypoints.utils._materialize_object_storage_configs",
-            return_value=str(tmp_path),
-        )
-        get_config = mocker.patch(
-            "vllm_omni.entrypoints.utils.get_config",
-            return_value=hf_config,
-        )
-
-        result = resolve_model_config_path(uri)
-
-        materialize.assert_called_once_with(uri)
-        get_config.assert_called_once()
-        assert get_config.call_args[0][0] == str(tmp_path)
-        assert result is not None
-        assert Path(result).as_posix().endswith("qwen3_omni_moe.yaml")
-
-
-class TestTryResolveOmniModelType:
-    """Name matching must only scan the last path component."""
-
-    def test_uri_bucket_name_cannot_hijack_pipeline(self):
-        # "qwen3-tts-models" normalizes to "qwen3ttsmodels", which contains the
-        # registered key "qwen3tts" — only the basename may participate, so a
-        # misleadingly named bucket selects nothing.
-        assert _try_resolve_omni_model_type("s3://qwen3-tts-models/plain-checkpoint") is None
-
-    def test_basename_still_matches_pipeline(self):
-        assert _try_resolve_omni_model_type("gs://any-bucket/Fun-CosyVoice3-0.5B-2512") == "cosyvoice3"
-

 class TestLoadAndResolveStageConfigs:
    def test_load_and_resolve_with_kwargs(self, mocker: MockerFixture):
@@ -11,16 +11,11 @@ affect in-process Omni (``OmniRunner`` / offline e2e) vs subprocess-based
 import pytest


-def skip_if_gated_repo_inaccessible(
-    repo_id: str,
-    *,
-    revision: str | None = None,
-    filename: str = "config.json",
-) -> None:
+def skip_if_gated_repo_inaccessible(repo_id: str) -> None:
    """Skip the test if a gated HuggingFace repo is not accessible.

-    Tries to download ``filename`` via ``hf_hub_download``, which performs an
-    actual file-access check (unlike ``HfApi().model_info()``
+    Tries to download the model's config.json via ``hf_hub_download``,
+    which performs an actual file-access check (unlike ``HfApi().model_info()``
    that only checks metadata).  If the token has metadata access but not
    file-download access, ``hf_hub_download`` will raise ``GatedRepoError``
    and we skip cleanly.
@@ -31,7 +26,7 @@ def skip_if_gated_repo_inaccessible(
    except Exception:
        return
    try:
-        hf_hub_download(repo_id=repo_id, filename=filename, revision=revision)
+        hf_hub_download(repo_id=repo_id, filename="config.json")
    except GatedRepoError as exc:
        pytest.skip(
            f"Skipping: gated HF repo {repo_id!r} inaccessible to the current "
@@ -1,363 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""CPU coverage for Gepard's prompt layout, streaming-decode window and config.
-
-The e2e test needs a GPU, NeMo and the checkpoint, so it only runs nightly.
-These pin the parts that are pure logic and break silently: which frames each
-decode covers, which samples survive the lookback trim, the prompt layout, and
-that the packaged deploy config sits where the example looks for it.
-
-The codec is replaced by a shift-invariant stand-in whose output depends only
-on the codes, which makes "chunked emissions concatenate to one whole-history
-decode" an exact assertion — the contract lookback exists to provide.
-"""
-
-from __future__ import annotations
-
-from pathlib import Path
-
-import pytest
-import torch
-import torch.nn as nn
-
-import vllm_omni
-from vllm_omni.config.stage_config import load_deploy_config
-from vllm_omni.model_executor.models.gepard.configuration_gepard import GepardConfig
-from vllm_omni.model_executor.models.gepard.gepard_talker import (
-    CHUNK_FRAMES,
-    FIRST_CHUNK_FRAMES,
-    GepardTalkerForConditionalGeneration,
-    _GepardState,
-)
-from vllm_omni.model_executor.models.gepard.nanocodec import (
-    LOOKBACK_FRAMES,
-    NanoCodec,
-    apply_end_of_speech_tail,
-)
-from vllm_omni.model_executor.models.gepard.prompt import build_gepard_prompt_ids
-
-pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
-
-SAMPLES_PER_FRAME = 1024
-NUM_HEADS = 32
-SAMPLE_RATE = 22050
-# Text-token count at or above which the layout is left unrepeated, read from
-# the checkpoint defaults so the boundary cases follow the config.
-APPLY_BELOW = GepardConfig().text_repetition_apply_below
-
-
-class _ShiftInvariantCodec:
-    """Frame tagged ``v`` -> samples ``v*SPF .. v*SPF+SPF-1``, wherever it sits."""
-
-    def __init__(self) -> None:
-        self.windows: list[int] = []  # frame count of every decode call
-
-    def decode_from_codes(self, codes: torch.Tensor, codes_len: torch.Tensor):
-        assert codes.shape[0] == 1, "streaming decode is per-request"
-        assert codes.shape[1] == NUM_HEADS, f"expected {NUM_HEADS} code channels"
-        tags = codes[0, 0, :].to(torch.int64)  # (T,)
-        self.windows.append(int(tags.shape[0]))
-        offsets = torch.arange(SAMPLES_PER_FRAME, dtype=torch.int64)
-        audio = (tags.unsqueeze(1) * SAMPLES_PER_FRAME + offsets).reshape(1, -1)
-        return audio.to(torch.float32), codes_len * SAMPLES_PER_FRAME
-
-
-def _make_codec() -> tuple[NanoCodec, _ShiftInvariantCodec]:
-    codec = NanoCodec(codec_id="test/fake-codec", sample_rate=SAMPLE_RATE)
-    fake = _ShiftInvariantCodec()
-    codec._codec = fake  # bypasses load(); NeMo is never imported
-    assert codec.is_loaded
-    return codec, fake
-
-
-def _make_talker(codec: NanoCodec) -> GepardTalkerForConditionalGeneration:
-    """A real talker with only the audio-path attributes populated.
-
-    ``__init__`` needs a VllmConfig and builds the backbone; the methods under
-    test touch only these four attributes.
-    """
-    talker = GepardTalkerForConditionalGeneration.__new__(GepardTalkerForConditionalGeneration)
-    nn.Module.__init__(talker)  # sets up _parameters/_buffers/_modules only
-    talker._codec = codec
-    talker._audio_queue = []
-    talker._active_states = {}
-    talker._deferred_cleanup_ids = set()
-    return talker
-
-
-def _frame(tag: int) -> torch.Tensor:
-    """One committed frame: 32 per-dimension codes, all carrying the tag."""
-    return torch.full((NUM_HEADS,), tag, dtype=torch.long)
-
-
-def _expected_samples(num_frames: int) -> torch.Tensor:
-    offsets = torch.arange(SAMPLES_PER_FRAME, dtype=torch.int64)
-    tags = torch.arange(num_frames, dtype=torch.int64).unsqueeze(1)
-    return (tags * SAMPLES_PER_FRAME + offsets).reshape(-1).to(torch.float32)
-
-
-def _drain(talker) -> torch.Tensor:
-    """Concatenate and clear the queued deltas, as make_omni_output does."""
-    if not talker._audio_queue:
-        return torch.empty(0)
-    out = torch.cat([a.reshape(-1) for _, a in talker._audio_queue])
-    talker._audio_queue.clear()
-    return out
-
-
-def _generate(talker, state, num_frames: int, *, stop_at_end: bool) -> None:
-    """Drive the commit loop the way forward() does, one frame per step."""
-    for tag in range(num_frames):
-        state.frames.append(_frame(tag))
-        state.frame_count += 1
-        state.past_first_step = True
-        talker._emit_audio(state, is_final=stop_at_end and tag == num_frames - 1)
-
-
-# Counts straddling both cadence thresholds, where an off-by-one hides.
-@pytest.mark.parametrize(
-    "num_frames",
-    [
-        1,
-        2,
-        FIRST_CHUNK_FRAMES - 1,
-        FIRST_CHUNK_FRAMES,
-        FIRST_CHUNK_FRAMES + 1,
-        CHUNK_FRAMES - 1,
-        CHUNK_FRAMES,
-        CHUNK_FRAMES + 1,
-        FIRST_CHUNK_FRAMES + CHUNK_FRAMES,
-        79,
-        200,
-    ],
-)
-def test_chunked_emissions_equal_a_whole_history_decode(num_frames: int) -> None:
-    codec, _fake = _make_codec()
-    talker = _make_talker(codec)
-    state = _GepardState(request_id="req-0")
-
-    _generate(talker, state, num_frames, stop_at_end=True)
-
-    torch.testing.assert_close(_drain(talker), _expected_samples(num_frames))
-    assert state.emitted_frames == num_frames
-
-
-def test_every_frame_is_emitted_exactly_once() -> None:
-    """The trim must not drop or duplicate samples at a chunk boundary."""
-    codec, _fake = _make_codec()
-    talker = _make_talker(codec)
-    state = _GepardState(request_id="req-0")
-    num_frames = FIRST_CHUNK_FRAMES + 2 * CHUNK_FRAMES + 3
-
-    _generate(talker, state, num_frames, stop_at_end=True)
-    audio = _drain(talker)
-
-    assert audio.numel() == num_frames * SAMPLES_PER_FRAME
-    # Tags are globally unique, so a duplicated or dropped frame shows up here.
-    assert len(torch.unique(audio)) == audio.numel()
-
-
-def test_decode_windows_carry_lookback_context() -> None:
-    """Every decode after the first must re-decode LOOKBACK_FRAMES of context."""
-    codec, fake = _make_codec()
-    talker = _make_talker(codec)
-    state = _GepardState(request_id="req-0")
-    num_frames = FIRST_CHUNK_FRAMES + 2 * CHUNK_FRAMES
-
-    _generate(talker, state, num_frames, stop_at_end=True)
-
-    assert len(fake.windows) >= 3, "expected a first chunk plus steady-state chunks"
-    assert fake.windows[0] == FIRST_CHUNK_FRAMES, "the first chunk has no left context"
-    for window in fake.windows[1:]:
-        assert window == CHUNK_FRAMES + LOOKBACK_FRAMES
-
-
-def test_codec_runs_once_per_chunk_not_once_per_frame() -> None:
-    codec, fake = _make_codec()
-    talker = _make_talker(codec)
-    state = _GepardState(request_id="req-0")
-    num_frames = FIRST_CHUNK_FRAMES + 3 * CHUNK_FRAMES
-
-    _generate(talker, state, num_frames, stop_at_end=True)
-
-    assert len(fake.windows) == 4
-    assert sum(fake.windows) < num_frames * 2  # far below one decode per frame
-
-
-def test_final_emit_drains_a_partial_chunk() -> None:
-    """A request can end mid-chunk (max_tokens, STOP); the frames committed
-    since the last cadence boundary must still reach the queue.
-
-    Arithmetic only: this drives ``_emit_audio`` directly, so it says nothing
-    about *whether* the engine reaches it on the request's last step. That
-    wiring lives in ``test_gepard_wiring.py`` — a bug there leaves this green.
-    """
-    codec, _fake = _make_codec()
-    talker = _make_talker(codec)
-    state = _GepardState(request_id="req-0")
-    num_frames = FIRST_CHUNK_FRAMES + CHUNK_FRAMES + 5
-
-    _generate(talker, state, num_frames, stop_at_end=False)
-    assert state.emitted_frames < num_frames, "a partial chunk should still be pending"
-
-    talker._emit_audio(state, is_final=True)
-
-    torch.testing.assert_close(_drain(talker), _expected_samples(num_frames))
-    assert state.emitted_frames == num_frames
-
-
-def test_emitting_with_nothing_pending_is_a_no_op() -> None:
-    codec, fake = _make_codec()
-    talker = _make_talker(codec)
-    state = _GepardState(request_id="req-0")
-
-    _generate(talker, state, 4, stop_at_end=True)
-    calls_after_flush = len(fake.windows)
-    talker._audio_queue.clear()
-
-    talker._emit_audio(state, is_final=True)
-
-    assert len(fake.windows) == calls_after_flush, "codec ran with no new frames"
-    assert talker._audio_queue == []
-
-
-def test_concurrent_requests_keep_separate_windows() -> None:
-    codec, _fake = _make_codec()
-    talker = _make_talker(codec)
-    states = [_GepardState(request_id=f"req-{i}") for i in range(2)]
-
-    for state in states:
-        _generate(talker, state, FIRST_CHUNK_FRAMES, stop_at_end=True)
-
-    by_req: dict[str, list[torch.Tensor]] = {}
-    for req_id, audio in talker._audio_queue:
-        by_req.setdefault(req_id, []).append(audio.reshape(-1))
-    assert set(by_req) == {"req-0", "req-1"}
-    for chunks in by_req.values():
-        torch.testing.assert_close(torch.cat(chunks), _expected_samples(FIRST_CHUNK_FRAMES))
-
-
-def test_end_of_speech_tail_fades_and_pads() -> None:
-    audio = torch.ones(SAMPLE_RATE)  # 1 s of full-scale signal
-    out = apply_end_of_speech_tail(audio, SAMPLE_RATE, fade_ms=10.0, silence_ms=20.0)
-
-    fade = int(SAMPLE_RATE * 10.0 / 1000.0)
-    pad = int(SAMPLE_RATE * 20.0 / 1000.0)
-    assert out.numel() == SAMPLE_RATE + pad
-    assert torch.all(out[-pad:] == 0.0), "trailing pad must be silent"
-    assert out[SAMPLE_RATE - fade - 1] == 1.0, "the fade must not reach back past fade_ms"
-    # Monotonically ramped to zero — the pad alone would leave the click.
-    ramp = out[SAMPLE_RATE - fade : SAMPLE_RATE]
-    assert torch.all(ramp[:-1] >= ramp[1:])
-    assert ramp[-1] == 0.0
-
-
-def test_end_of_speech_tail_is_a_no_op_by_default() -> None:
-    audio = torch.ones(128)
-    out = apply_end_of_speech_tail(audio, SAMPLE_RATE, fade_ms=0.0, silence_ms=0.0)
-    assert out is audio
-
-
-def _repeat_count(ids: list[int], cfg: GepardConfig) -> int:
-    """How many ``[SOT ... EOT]`` copies the layout carries."""
-    return ids.count(cfg.start_of_text)
-
-
-def test_gepard_prompt_layout() -> None:
-    """The prompt carries the speaker slots and exactly one SOS, at the end."""
-    cfg = GepardConfig()
-    ids = build_gepard_prompt_ids([1, 2, 3], config=cfg)
-
-    slots = [cfg.speaker_token_base + i for i in range(cfg.num_speaker_prefix)]
-    assert ids[: cfg.num_speaker_prefix] == slots
-    assert ids.count(cfg.start_of_speech) == 1, "only the canonical copy may carry SOS"
-    assert ids[-1] == cfg.start_of_speech, "the first frame is sampled from the SOS position"
-
-
-def test_short_text_repeats_to_reach_the_target_token_mass() -> None:
-    """A short text is repeated until its region holds ~target_text_tokens."""
-    cfg = GepardConfig()
-    text = [1, 2, 3, 4]  # target 16 / 4 -> 4 copies
-    ids = build_gepard_prompt_ids(text, config=cfg)
-
-    assert _repeat_count(ids, cfg) == 4
-    assert ids.count(cfg.start_of_speech) == 1, "only the last copy opens the speech region"
-    assert ids[-1] == cfg.start_of_speech
-
-
-@pytest.mark.parametrize(
-    ("num_text_tokens", "repeated"),
-    [
-        (APPLY_BELOW - 1, True),
-        (APPLY_BELOW, False),
-        (APPLY_BELOW + 1, False),
-    ],
-)
-def test_apply_below_is_the_repetition_boundary(num_text_tokens: int, repeated: bool) -> None:
-    """Both sides of the threshold: at apply_below the text is left alone.
-
-    The layout has to match training here or WER collapses, and the boundary
-    is where an off-by-one would be invisible in a single happy-path clip.
-    """
-    cfg = GepardConfig()
-    ids = build_gepard_prompt_ids(list(range(1, num_text_tokens + 1)), config=cfg)
-
-    assert (_repeat_count(ids, cfg) > 1) is repeated
-
-
-def test_max_repeats_caps_a_one_token_text() -> None:
-    """target/1 would ask for 16 copies; the cap keeps the prefill bounded."""
-    cfg = GepardConfig()
-    ids = build_gepard_prompt_ids([1], config=cfg)
-
-    assert _repeat_count(ids, cfg) == cfg.text_repetition_max_repeats
-
-
-def test_repetition_thresholds_are_read_from_the_config() -> None:
-    """The numbers come from the checkpoint sidecar, not from literals.
-
-    A checkpoint whose text_repetition block differs must change the layout;
-    duplicating the values in the builder would silently ignore it.
-    """
-    cfg = GepardConfig(text_repetition={"enabled": True, "target_text_tokens": 9, "apply_below": 8, "max_repeats": 2})
-    assert _repeat_count(build_gepard_prompt_ids([1, 2, 3], config=cfg), cfg) == 2  # ceil(9/3)=3, capped at 2
-    assert _repeat_count(build_gepard_prompt_ids(list(range(8)), config=cfg), cfg) == 1  # at apply_below
-
-
-def test_repetition_can_be_disabled() -> None:
-    cfg = GepardConfig(text_repetition={"enabled": False})
-    ids = build_gepard_prompt_ids([1], config=cfg)
-
-    assert _repeat_count(ids, cfg) == 1
-    assert ids[-1] == cfg.start_of_speech
-
-
-def test_empty_text_is_rejected() -> None:
-    """An empty text would still assemble a valid layout and be voiced.
-
-    ``[slots, SOT, EOT, SOS]`` parses, prefills and generates: the failure is
-    audible, not raised, which is the worst shape for a caller to debug.
-    """
-    with pytest.raises(ValueError, match="at least one text token"):
-        build_gepard_prompt_ids([], config=GepardConfig())
-
-
-def test_packaged_deploy_config_is_where_the_example_looks_for_it() -> None:
-    """The example resolves its default off ``vllm_omni.__file__``; tests reach
-    the same YAML through ``tests.helpers.stage_config``, so nothing covered the
-    packaged location until this.
-
-    ``pipeline: gepard`` is the load-bearing key: the checkpoint self-identifies
-    as qwen3_5_text, so without the pin the architectures fallback routes it to
-    the diffusion registry.
-
-    This asserts the location, not the example's constant — importing the
-    example would pull ``Omni`` and NeMo into a CPU test. Running the quick
-    start verbatim stays a release check.
-    """
-    path = Path(vllm_omni.__file__).resolve().parent / "deploy" / "gepard.yaml"
-    assert path.is_file(), f"packaged deploy config missing at {path}"
-
-    deploy = load_deploy_config(path)
-    assert deploy.pipeline == "gepard"
-    assert deploy.stages, "deploy config declares no stages"
@@ -1,731 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""CPU coverage for the seam between Gepard's talker and the AR runner.
-
-``test_gepard_window.py`` pins the decode arithmetic *below* this layer, and
-the e2e test exercises the layer *above* it but only along the path where the
-stop head fires. Neither covers the wiring itself: whether anything calls the
-final emit on a request's last step, and whether the payload it produces
-survives the runner's sparse-audio routing.
-
-These drive the real ``preprocess`` -> ``forward`` -> ``make_omni_output``
-sequence through a stand-in for ``GPUARModelRunner.execute_model``, and use the
-runner's own ``_resolve_sparse_mm_routing`` to decide what actually reaches the
-caller. Only the LM backbone and the frame sampler are stubbed out; every
-method under test is the real one, so a test here goes red when the engine
-path breaks — which a test that calls the flush by hand cannot do.
-"""
-
-from __future__ import annotations
-
-import logging
-
-import pytest
-import torch
-import torch.nn as nn
-
-from vllm_omni.model_executor.models.gepard.configuration_gepard import GepardConfig
-from vllm_omni.model_executor.models.gepard.gepard_talker import (
-    CHUNK_FRAMES,
-    FIRST_CHUNK_FRAMES,
-    GepardTalkerForConditionalGeneration,
-    _ReqInfo,
-)
-from vllm_omni.model_executor.models.gepard.nanocodec import NanoCodec
-from vllm_omni.model_executor.models.gepard.prompt import build_gepard_prompt_ids
-from vllm_omni.worker.gpu_ar_model_runner import GPUARModelRunner
-
-pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
-
-SAMPLES_PER_FRAME = 1024
-NUM_HEADS = 32
-SAMPLE_RATE = 22050
-LOGGER_NAME = "vllm_omni.model_executor.models.gepard.gepard_talker"
-
-# Each request tags its frames from a distinct base, so a sample that reaches
-# the wrong request is visible in the value, not just the length.
-_TAG_STRIDE = 100_000
-
-
-class _TagCodec:
-    """Frame carrying tag ``v`` -> samples ``v*SPF .. v*SPF+SPF-1``, wherever it sits.
-
-    Shift-invariant, so "what the caller received" can be compared against one
-    whole-history decode exactly. The tag rides channel 1: channel 0 is head0,
-    which the talker keeps inside its real (small) vocab.
-    """
-
-    def __init__(self) -> None:
-        self.decode_calls = 0
-
-    def decode_from_codes(self, codes: torch.Tensor, codes_len: torch.Tensor):
-        assert codes.shape[0] == 1, "streaming decode is per-request"
-        assert codes.shape[1] == NUM_HEADS, f"expected {NUM_HEADS} code channels"
-        self.decode_calls += 1
-        tags = codes[0, 1, :].to(torch.int64)  # (T,)
-        offsets = torch.arange(SAMPLES_PER_FRAME, dtype=torch.int64)
-        audio = (tags.unsqueeze(1) * SAMPLES_PER_FRAME + offsets).reshape(1, -1)
-        return audio.to(torch.float32), codes_len * SAMPLES_PER_FRAME
-
-
-class _StubBackbone(nn.Module):
-    """Stands in for the Qwen3.5 backbone; only shapes matter to this seam."""
-
-    def __init__(self, hidden: int) -> None:
-        super().__init__()
-        self.hidden = hidden
-
-    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
-        return torch.zeros(int(input_ids.shape[0]), self.hidden)
-
-    def forward(
-        self,
-        input_ids: torch.Tensor | None = None,
-        positions: torch.Tensor | None = None,
-        intermediate_tensors=None,
-        inputs_embeds: torch.Tensor | None = None,
-    ) -> torch.Tensor:
-        src = inputs_embeds if inputs_embeds is not None else input_ids
-        return torch.zeros(int(src.shape[0]), self.hidden)
-
-
-def _expected_samples(tag_base: int, num_frames: int) -> torch.Tensor:
-    offsets = torch.arange(SAMPLES_PER_FRAME, dtype=torch.int64)
-    tags = (torch.arange(num_frames, dtype=torch.int64) + tag_base).unsqueeze(1)
-    return (tags * SAMPLES_PER_FRAME + offsets).reshape(-1).to(torch.float32)
-
-
-class _MiniEngine:
-    """Drives the talker the way ``GPUARModelRunner.execute_model`` does.
-
-    Reproduces the three orderings the dropped-tail bug lived in:
-
-    * ``on_requests_finished`` fires at the START of the step AFTER the one
-      that finished the request (it reads ``scheduler_output.finished_req_ids``);
-    * a finished request is not scheduled again, so it is absent from
-      ``req_ids_output_copy`` and from that step's routing;
-    * with nothing left to schedule there is no step at all, so no further
-      ``forward`` runs;
-    * a stop-token finish gets one more scheduled step, a budget finish does
-      not -- under async scheduling only the latter is guarded against it.
-    """
-
-    def __init__(self, talker: GepardTalkerForConditionalGeneration, *, max_model_len: int) -> None:
-        self.talker = talker
-        self.max_model_len = max_model_len
-        self.reqs: dict[str, dict] = {}
-        self.order: list[str] = []
-        self._finished_last_step: list[str] = []
-        self.delivered: dict[str, list[torch.Tensor]] = {}
-        self.committed_frames: dict[str, int] = {}
-        self.steps = 0
-        self._tags_handed_out = 0
-
-    def add_request(
-        self,
-        req_id: str,
-        *,
-        max_tokens: int,
-        stop_after: int | None = None,
-        text_len: int = 3,
-        seed: int | None = None,
-    ) -> None:
-        """Submit a request. An id already seen is a resubmit of that id.
-
-        Tags come from a counter, so a resubmitted id gets frames the previous
-        request could not have produced.
-        """
-        prompt_ids = build_gepard_prompt_ids(list(range(10, 10 + text_len)), config=self.talker.config)
-        self._tags_handed_out += 1
-        self.reqs[req_id] = {
-            "prompt_ids": torch.tensor(prompt_ids, dtype=torch.long),
-            "prompt_len": len(prompt_ids),
-            "max_tokens": max_tokens,
-            "stop_after": stop_after,
-            "stop_fired": False,
-            "seed": seed,
-            "computed": 0,
-            "generated": 0,
-            "finished": False,
-            "ghost_step_pending": False,
-            "next_token": None,
-            "tag_base": _TAG_STRIDE * self._tags_handed_out,
-        }
-        if req_id not in self.order:
-            self.order.append(req_id)
-        self.delivered[req_id] = []
-        self.committed_frames[req_id] = 0
-
-    def abort(self, req_id: str) -> None:
-        """Kill a request the way an aborted client does: no STOP, no budget."""
-        self.reqs[req_id]["finished"] = True
-        self._finished_last_step.append(req_id)
-
-    def _live(self) -> list[str]:
-        return [r for r in self.order if not self.reqs[r]["finished"]]
-
-    def step(self) -> bool:
-        """Run one engine step. False when there was nothing to schedule."""
-        live = self._live()
-        if not live:
-            return False
-        talker = self.talker
-        self.steps += 1
-
-        # --- runner: notify the model of requests finished since last step ---
-        if self._finished_last_step:
-            talker.on_requests_finished(set(self._finished_last_step))
-            self._finished_last_step = []
-
-        # --- runner: per-request preprocess, in input_batch.req_ids order ---
-        embeds_parts, ids_parts = [], []
-        for req_id in live:
-            r = self.reqs[req_id]
-            is_prefill = r["computed"] < r["prompt_len"]
-            if is_prefill:
-                ids = r["prompt_ids"]
-            else:
-                ids = torch.tensor([r["next_token"] or 0], dtype=torch.long)
-            span = int(ids.shape[0])
-            out_ids, out_embeds, _ = talker.preprocess(
-                input_ids=ids,
-                input_embeds=None,
-                request_id=req_id,
-                _omni_prompt_len=r["prompt_len"],
-                _omni_num_computed_tokens=r["computed"],
-                _omni_is_prefill=is_prefill,
-                _omni_max_tokens=r["max_tokens"],
-                _omni_seed=r["seed"],
-            )
-            ids_parts.append(out_ids)
-            embeds_parts.append(out_embeds)
-            r["span"] = span
-
-        inputs_embeds = torch.cat(embeds_parts, dim=0)
-        input_ids = torch.cat(ids_parts, dim=0)
-        positions = torch.arange(int(input_ids.shape[0]), dtype=torch.long)
-
-        # --- runner: forward, then build the multimodal payload ---
-        hidden = talker.forward(input_ids, positions, inputs_embeds=inputs_embeds)
-        results = list(talker._results_queue)
-        omni_out = talker.make_omni_output(hidden)
-        mm = omni_out.multimodal_outputs
-        # compute_logits drains this; it does not touch the audio path.
-        talker._results_queue.clear()
-
-        for req_id in live:
-            state = talker._active_states.get(req_id)
-            if state is not None:
-                self.committed_frames[req_id] = state.frame_count
-
-        # --- runner: sparse-audio routing decides what reaches the caller ---
-        downstream, sparse_index, is_sparse = GPUARModelRunner._resolve_sparse_mm_routing(
-            engine_output_type="audio",
-            req_ids_output_copy=list(live),
-            downstream_req_ids=list(live),
-            multimodal_outputs=mm,
-        )
-        assert is_sparse, "gepard always marks meta.sparse_audio"
-        for rid in downstream:
-            self.delivered[rid].append(mm["model_outputs"][sparse_index[rid]].reshape(-1))
-
-        # --- runner/scheduler bookkeeping: advance and detect finishes ---
-        stopped = {rid for rid, _head0, do_stop in results if do_stop}
-        for req_id in live:
-            r = self.reqs[req_id]
-            r["computed"] += r["span"]
-            r["generated"] = r["computed"] - r["prompt_len"] + 1
-            r["next_token"] = 0
-            length_cap = min(r["max_tokens"], self.max_model_len - r["prompt_len"])
-            if r["ghost_step_pending"]:
-                # This step WAS the extra one. The engine finished the request
-                # on the stop token last step and discards whatever came back
-                # here, whether or not the model chose to stop again.
-                r["finished"] = True
-                self._finished_last_step.append(req_id)
-            elif req_id in stopped:
-                # Ends on the stop token: one more step is already scheduled.
-                r["ghost_step_pending"] = True
-            elif r["generated"] >= length_cap:
-                r["finished"] = True
-                self._finished_last_step.append(req_id)
-        return True
-
-    def run(self, max_steps: int = 4000) -> None:
-        for _ in range(max_steps):
-            if not self.step():
-                return
-        raise AssertionError("engine did not drain")
-
-    def audio_for(self, req_id: str) -> torch.Tensor:
-        chunks = self.delivered[req_id]
-        return torch.cat(chunks) if chunks else torch.empty(0)
-
-
-def _make_talker(*, max_model_len: int = 4096) -> tuple[GepardTalkerForConditionalGeneration, _TagCodec]:
-    """A real talker with only the audio-path collaborators populated.
-
-    ``__init__`` needs a VllmConfig and builds the Qwen3.5 backbone; the seam
-    under test needs none of that, so the backbone and the frame sampler are
-    stubbed and everything else is the class's own code.
-    """
-    cfg = GepardConfig()
-    # Fixed, not cfg.get_text_config().hidden_size: nothing in this seam reads
-    # the real width, and a small one keeps the test off the backbone config.
-    hidden = 64
-    codec = NanoCodec(codec_id="test/fake-codec", sample_rate=SAMPLE_RATE)
-    fake = _TagCodec()
-    codec._codec = fake  # bypasses load(); NeMo is never imported
-    assert codec.is_loaded
-
-    talker = GepardTalkerForConditionalGeneration.__new__(GepardTalkerForConditionalGeneration)
-    nn.Module.__init__(talker)
-    talker.config = cfg
-    talker.model = _StubBackbone(hidden)
-    # requires_grad=False: preprocess writes the prefix into inputs_embeds
-    # in place, and production runs that under inference_mode.
-    talker.null_prefix = nn.Parameter(torch.zeros(cfg.num_speaker_prefix, hidden), requires_grad=False)
-    talker._codec = codec
-    talker._sample_rate = SAMPLE_RATE
-    talker._max_model_len = max_model_len
-    talker._active_states = {}
-    talker._pending_requests = []
-    talker._results_queue = []
-    talker._audio_queue = []
-    talker._deferred_cleanup_ids = set()
-    return talker, fake
-
-
-def _install_scripted_sampling(talker: GepardTalkerForConditionalGeneration, engine: _MiniEngine) -> None:
-    """Replace the 32-head sampler with a scripted one.
-
-    Frames carry a per-request, per-step tag so the caller's audio is
-    self-describing. STOP fires once, on the request's configured frame: on the
-    extra step the engine then runs, re-firing is likely but not guaranteed,
-    and the model may not assume it.
-    """
-    head0_vocab = talker.config.head0_vocab_size
-
-    # Signature mirrors the real ``_sample_frame`` minus ``self``; the scripted
-    # sampler ignores ``generators`` because seeding is decided in
-    # ``_build_with_real_sampler``, which keeps the model's own sampler.
-    def _sample_frame(rows: torch.Tensor, generators: list[torch.Generator | None] | None = None):
-        n = int(rows.shape[0])
-        head0 = torch.zeros(n, dtype=torch.long)
-        heads = torch.zeros(n, NUM_HEADS - 1, dtype=torch.long)
-        stop = torch.zeros(n, dtype=torch.bool)
-        for i, (req_id, _span, _samples_frame, *_rest) in enumerate(talker._pending_requests):
-            state = talker._active_states.get(req_id)
-            frame_index = state.frame_count if state is not None else 0
-            tag = engine.reqs[req_id]["tag_base"] + frame_index
-            head0[i] = tag % head0_vocab
-            heads[i] = tag
-            r = engine.reqs[req_id]
-            if r["stop_after"] is not None and frame_index >= r["stop_after"] and not r["stop_fired"]:
-                r["stop_fired"] = True
-                stop[i] = True
-        return head0, heads, stop
-
-    talker._sample_frame = _sample_frame
-    talker._audio_frame_embed = lambda h0, h: torch.zeros(int(h0.shape[0]), talker.model.hidden)
-
-
-def _build(*, max_model_len: int = 4096) -> tuple[GepardTalkerForConditionalGeneration, _TagCodec, _MiniEngine]:
-    talker, codec = _make_talker(max_model_len=max_model_len)
-    engine = _MiniEngine(talker, max_model_len=max_model_len)
-    _install_scripted_sampling(talker, engine)
-    return talker, codec, engine
-
-
-def _build_with_real_sampler(*, max_model_len: int = 4096) -> tuple[GepardTalkerForConditionalGeneration, _MiniEngine]:
-    """Same seam, but ``_sample_frame`` is the model's own 32-head sampler.
-
-    The stub backbone hands every row the same hidden state, so the sampled
-    codes — and therefore the audio the codec tags them into — vary with
-    nothing but the Gumbel draws. That makes this the one place where "did the
-    caller's seed reach the audio" is a decidable question.
-    """
-    talker, _codec = _make_talker(max_model_len=max_model_len)
-    engine = _MiniEngine(talker, max_model_len=max_model_len)
-    cfg = talker.config
-    talker.vocab_sizes = list(cfg.audio_head_levels)
-    talker.num_heads = cfg.num_audio_heads
-    talker.head0_vocab = cfg.head0_vocab_size
-    talker.stop_token = cfg.stop_token
-    talker.stop_threshold = cfg.stop_threshold
-    talker.temperature = cfg.temperature
-    talker._greedy = False
-    gather_idx, mask = GepardTalkerForConditionalGeneration._build_gather_mask(talker.vocab_sizes)
-    talker._cb_gather_idx = gather_idx
-    talker._cb_mask = mask
-    # Fixed weights: two engine runs must differ only by their RNG.
-    torch.manual_seed(0)
-    talker.fused_codebook_head = nn.Linear(talker.model.hidden, sum(talker.vocab_sizes))
-    talker.stop_head = nn.Linear(talker.model.hidden, 1)
-    # Never stop: these requests end on their token budget, so the frame count
-    # is fixed and only the sampled codes can differ.
-    torch.nn.init.constant_(talker.stop_head.bias, -20.0)
-    talker._audio_frame_embed = lambda h0, h: torch.zeros(int(h0.shape[0]), talker.model.hidden)
-    return talker, engine
-
-
-# --------------------------------------------------------------------------
-# The tail must reach the caller, whichever way the request ends.
-# --------------------------------------------------------------------------
-
-
-def test_last_in_flight_request_delivers_every_committed_frame() -> None:
-    """The only request runs out of budget: there is no later step to flush in.
-
-    This is the case a deferred flush cannot serve at all — nothing runs after
-    the final forward.
-    """
-    _talker, _codec, engine = _build()
-    engine.add_request("req-0", max_tokens=FIRST_CHUNK_FRAMES + CHUNK_FRAMES + 5)
-    engine.run()
-
-    frames = engine.committed_frames["req-0"]
-    assert frames > 0
-    torch.testing.assert_close(
-        engine.audio_for("req-0"),
-        _expected_samples(engine.reqs["req-0"]["tag_base"], frames),
-    )
-
-
-@pytest.mark.parametrize(
-    "max_tokens",
-    [
-        1,
-        FIRST_CHUNK_FRAMES,
-        FIRST_CHUNK_FRAMES + 1,
-        30,
-        FIRST_CHUNK_FRAMES + CHUNK_FRAMES,
-        100,
-        150,
-    ],
-)
-def test_budget_truncated_request_delivers_every_committed_frame(max_tokens: int) -> None:
-    """Lowering max_tokens must shorten the clip, never silently cut it back
-    to the last cadence boundary."""
-    _talker, _codec, engine = _build()
-    engine.add_request("req-0", max_tokens=max_tokens)
-    engine.run()
-
-    frames = engine.committed_frames["req-0"]
-    assert frames == max_tokens, "every sampled step commits one frame when STOP never fires"
-    torch.testing.assert_close(
-        engine.audio_for("req-0"),
-        _expected_samples(engine.reqs["req-0"]["tag_base"], frames),
-    )
-
-
-def test_stop_head_path_still_delivers_every_committed_frame() -> None:
-    """Regression guard: the path that already worked must keep working."""
-    _talker, _codec, engine = _build()
-    engine.add_request("req-0", max_tokens=4000, stop_after=FIRST_CHUNK_FRAMES + 7)
-    engine.run()
-
-    frames = engine.committed_frames["req-0"]
-    assert frames == FIRST_CHUNK_FRAMES + 7, "the STOP frame itself is not committed"
-    torch.testing.assert_close(
-        engine.audio_for("req-0"),
-        _expected_samples(engine.reqs["req-0"]["tag_base"], frames),
-    )
-
-
-def test_the_step_after_a_stop_token_commits_nothing(caplog: pytest.LogCaptureFixture) -> None:
-    """Async scheduling runs one more step on a request that already stopped.
-
-    Committing its frame would strand it — nothing ships after the stop step —
-    and would make the truncation warning fire on a request that delivered
-    everything it produced, which is the one thing that warning must not do.
-    """
-    talker, _codec, engine = _build()
-    stop_after = FIRST_CHUNK_FRAMES + 6
-    engine.add_request("req-0", max_tokens=4000, stop_after=stop_after)
-
-    target = logging.getLogger(LOGGER_NAME)
-    target.addHandler(caplog.handler)
-    prev = target.level
-    target.setLevel(logging.WARNING)
-    try:
-        engine.run()
-        # The cleanup that would warn runs on the step after the request
-        # leaves, and here there is none. Drive it the way the runner would.
-        talker.on_requests_finished({"req-0"})
-        talker._flush_deferred_cleanup()
-    finally:
-        target.removeHandler(caplog.handler)
-        target.setLevel(prev)
-
-    assert engine.reqs["req-0"]["ghost_step_pending"], "the harness did not run the extra step"
-    frames = engine.committed_frames["req-0"]
-    assert frames == stop_after, f"the extra step committed a frame: {frames} > {stop_after}"
-    torch.testing.assert_close(
-        engine.audio_for("req-0"),
-        _expected_samples(engine.reqs["req-0"]["tag_base"], frames),
-    )
-    assert not [m for m in caplog.messages if "undelivered frame" in m], (
-        f"a fully delivered request was reported as truncated; saw {caplog.messages}"
-    )
-
-
-def test_a_resubmitted_request_id_keeps_its_new_state(caplog: pytest.LogCaptureFixture) -> None:
-    """An aborted id can come back before its cleanup has run.
-
-    The runner treats the two as distinct requests, so the pending free must
-    not take the second one's state: without that, the state is popped at the
-    end of the very forward that created it and every later decode falls back
-    to the zero-embedding path — the request keeps producing frames, so only
-    the audio is wrong.
-    """
-    _talker, _codec, engine = _build()
-    engine.add_request("req-0", max_tokens=4000)
-    for _ in range(FIRST_CHUNK_FRAMES + 2):
-        engine.step()
-    engine.abort("req-0")
-
-    # Same id, resubmitted before the model has been told about the abort.
-    engine.add_request("req-0", max_tokens=FIRST_CHUNK_FRAMES + CHUNK_FRAMES + 5)
-
-    target = logging.getLogger(LOGGER_NAME)
-    target.addHandler(caplog.handler)
-    prev = target.level
-    target.setLevel(logging.WARNING)
-    try:
-        engine.run()
-    finally:
-        target.removeHandler(caplog.handler)
-        target.setLevel(prev)
-
-    frames = engine.committed_frames["req-0"]
-    assert frames == engine.reqs["req-0"]["max_tokens"], "the second request lost frames"
-    torch.testing.assert_close(
-        engine.audio_for("req-0"),
-        _expected_samples(engine.reqs["req-0"]["tag_base"], frames),
-        msg="the second request did not receive its own audio",
-    )
-    # The abort still gets reported: the cleanup that would have said so is the
-    # one this cancelled.
-    assert any("resubmitted" in m for m in caplog.messages), (
-        f"the aborted request's truncation went unreported; saw {caplog.messages}"
-    )
-
-
-def test_a_preempted_request_is_reported_rather_than_resumed(caplog: pytest.LogCaptureFixture) -> None:
-    """Recompute preemption re-prefills a request that was mid-generation.
-
-    This model cannot survive that, and the point of the test is that it says
-    so. The recomputed ids are the prompt plus the head0 codes sampled so far,
-    which go through the text embedding table, and each frame's other 31 codes
-    are never in the token stream at all — there is nothing to stitch the two
-    halves back together with, so generation restarts from frame 0.
-
-    The harness re-prefills with the bare prompt, which understates the damage.
-    What is under test is the report and the reset, not the corruption, which
-    needs a real backbone to observe.
-    """
-    _talker, _codec, engine = _build()
-    engine.add_request("req-0", max_tokens=4000)
-    for _ in range(FIRST_CHUNK_FRAMES + 2):
-        engine.step()
-    assert engine.committed_frames["req-0"] > 0
-
-    # Preemption the way the scheduler does it: computed tokens back to 0, so
-    # the next step re-prefills, and no id reaches on_requests_finished.
-    engine.reqs["req-0"]["computed"] = 0
-
-    target = logging.getLogger(LOGGER_NAME)
-    target.addHandler(caplog.handler)
-    prev = target.level
-    target.setLevel(logging.WARNING)
-    try:
-        engine.step()
-    finally:
-        target.removeHandler(caplog.handler)
-        target.setLevel(prev)
-
-    assert any("preempted" in m for m in caplog.messages), (
-        f"a preempted request restarted its audio silently; saw {caplog.messages}"
-    )
-    assert not any("resubmitted" in m for m in caplog.messages), (
-        "a preemption was reported as an id resubmit; the two need different words"
-    )
-    assert engine.committed_frames["req-0"] == 1, "the re-prefill must start the frame history over"
-
-
-def test_concurrent_requests_each_receive_their_own_full_audio() -> None:
-    """One request finishes while another stays live.
-
-    The earlier finisher is the case where a deferred flush *does* run but the
-    payload is filtered out again, because its id has left the output batch.
-    """
-    _talker, _codec, engine = _build()
-    engine.add_request("req-short", max_tokens=FIRST_CHUNK_FRAMES + 3)
-    engine.add_request("req-long", max_tokens=FIRST_CHUNK_FRAMES + CHUNK_FRAMES + 11)
-    engine.run()
-
-    for req_id in ("req-short", "req-long"):
-        frames = engine.committed_frames[req_id]
-        assert frames == engine.reqs[req_id]["max_tokens"]
-        torch.testing.assert_close(
-            engine.audio_for(req_id),
-            _expected_samples(engine.reqs[req_id]["tag_base"], frames),
-            msg=f"{req_id} lost or gained samples",
-        )
-
-
-def test_codec_still_runs_once_per_chunk_not_once_per_frame() -> None:
-    """Emitting on the final step must not degrade into per-frame decoding."""
-    _talker, codec, engine = _build()
-    num_frames = FIRST_CHUNK_FRAMES + 3 * CHUNK_FRAMES
-    engine.add_request("req-0", max_tokens=num_frames)
-    engine.run()
-
-    assert codec.decode_calls == 4, f"expected one decode per chunk, got {codec.decode_calls}"
-
-
-# --------------------------------------------------------------------------
-# Deferred cleanup is cleanup: it must not try to produce output.
-# --------------------------------------------------------------------------
-
-
-def test_deferred_cleanup_never_queues_audio() -> None:
-    """Anything queued from the cleanup hook is unroutable by construction.
-
-    By the time it runs the id has left ``req_ids_output_copy``, so a payload
-    under it is dropped; queueing there only risks leaking one request's audio
-    into the next step's payload for a different request.
-    """
-    talker, _codec, engine = _build()
-    engine.add_request("req-0", max_tokens=FIRST_CHUNK_FRAMES + 4)
-    engine.add_request("req-1", max_tokens=FIRST_CHUNK_FRAMES + CHUNK_FRAMES)
-    while engine.step():
-        if engine.reqs["req-0"]["finished"] and not engine.reqs["req-1"]["finished"]:
-            break
-
-    talker._audio_queue.clear()
-    talker.on_requests_finished({"req-0"})
-    talker._flush_deferred_cleanup()
-
-    assert talker._audio_queue == [], "cleanup emitted audio that routing cannot deliver"
-    assert "req-0" not in talker._active_states, "cleanup must still free the state"
-
-
-def test_cleanup_warns_when_frames_are_left_undelivered(caplog: pytest.LogCaptureFixture) -> None:
-    """An abort is the one case with no last step to emit on. Say so."""
-    _talker, _codec, engine = _build()
-    engine.add_request("req-0", max_tokens=4000)
-    engine.add_request("req-1", max_tokens=4000)
-    for _ in range(FIRST_CHUNK_FRAMES + 3):
-        engine.step()
-    engine.abort("req-0")
-
-    target = logging.getLogger(LOGGER_NAME)
-    target.addHandler(caplog.handler)
-    prev = target.level
-    target.setLevel(logging.WARNING)
-    try:
-        engine.step()  # cleanup for req-0 runs at the start of this step
-    finally:
-        target.removeHandler(caplog.handler)
-        target.setLevel(prev)
-
-    assert any("undelivered frame" in m for m in caplog.messages), (
-        f"aborted request dropped audio without a warning; saw {caplog.messages}"
-    )
-
-
-# --------------------------------------------------------------------------
-# A request's seed must reach the in-model sampler.
-# --------------------------------------------------------------------------
-
-_SEEDED_FRAMES = FIRST_CHUNK_FRAMES + 5
-
-
-def _run_seeded(seed: int | None, *, neighbour: bool = False, neighbour_seed: int | None = None) -> torch.Tensor:
-    """Audio delivered to ``req-a``, optionally sharing the batch with another."""
-    _talker, engine = _build_with_real_sampler()
-    engine.add_request("req-a", max_tokens=_SEEDED_FRAMES, seed=seed)
-    if neighbour:
-        engine.add_request("req-b", max_tokens=_SEEDED_FRAMES, seed=neighbour_seed)
-    engine.run()
-    return engine.audio_for("req-a")
-
-
-def test_a_seeded_request_is_reproducible_within_one_process() -> None:
-    """The bug this pins: the 32 heads are sampled in ``forward``, so a seed
-    that only reaches vLLM's sampler leaves the audio at the mercy of the
-    global RNG — two identical requests in one process then diverge."""
-    torch.testing.assert_close(_run_seeded(1234), _run_seeded(1234))
-
-
-def test_different_seeds_produce_different_audio() -> None:
-    """Guards the opposite failure: a seed that is threaded but never used."""
-    assert not torch.equal(_run_seeded(1234), _run_seeded(4321))
-
-
-def test_a_seeded_request_is_unaffected_by_its_batch() -> None:
-    """Each seeded request draws its own noise, so the noise it sees cannot
-    depend on who else was scheduled alongside it.
-
-    The stub backbone is what makes that checkable. A real one is a bf16
-    batched matmul whose rows are not reproducible across batch shapes, so
-    end-to-end batch invariance is not attainable there; this pins the half
-    the model actually controls.
-    """
-    torch.testing.assert_close(_run_seeded(1234), _run_seeded(1234, neighbour=True, neighbour_seed=999))
-
-
-def test_an_unseeded_neighbour_does_not_disturb_a_seeded_request() -> None:
-    """The mixed case: one row draws from the global RNG, the other must not."""
-    torch.testing.assert_close(_run_seeded(1234), _run_seeded(1234, neighbour=True))
-
-
-# --------------------------------------------------------------------------
-# Budget arithmetic, straight through the helper.
-# --------------------------------------------------------------------------
-
-
-def _info(**kwargs) -> _ReqInfo:
-    base = {
-        "request_id": "req-0",
-        "is_prefill": False,
-        "prompt_len": 20,
-        "num_computed_tokens": 20,
-        "max_tokens": 10,
-        "seed": None,
-    }
-    base.update(kwargs)
-    return _ReqInfo(**base)
-
-
-def test_prefill_step_is_the_first_output_token() -> None:
-    talker, _codec, _engine = _build()
-    first = _info(is_prefill=True, num_computed_tokens=0, max_tokens=1)
-    assert talker._is_last_output_token(first, span=20, samples_frame=True) is True
-    more = _info(is_prefill=True, num_computed_tokens=0, max_tokens=2)
-    assert talker._is_last_output_token(more, span=20, samples_frame=True) is False
-
-
-def test_partial_prefill_chunk_does_not_consume_budget() -> None:
-    """vLLM discards that row's token, so it must not count against max_tokens."""
-    talker, _codec, _engine = _build()
-    info = _info(is_prefill=True, num_computed_tokens=0, prompt_len=40, max_tokens=1)
-    assert talker._is_last_output_token(info, span=10, samples_frame=False) is False
-
-
-def test_budget_is_capped_by_max_model_len() -> None:
-    """A max_tokens larger than the context window never gets reached."""
-    talker, _codec, _engine = _build(max_model_len=30)
-    # prompt 20 + 10 generated fills the window, whatever max_tokens says.
-    at_cap = _info(num_computed_tokens=28, max_tokens=1000)
-    assert talker._is_last_output_token(at_cap, span=1, samples_frame=True) is True
-    below_cap = _info(num_computed_tokens=27, max_tokens=1000)
-    assert talker._is_last_output_token(below_cap, span=1, samples_frame=True) is False
-
-
-def test_missing_budget_is_not_treated_as_exhausted() -> None:
-    """A runner that does not thread max_tokens must not trigger a fade."""
-    talker, _codec, _engine = _build(max_model_len=0)
-    assert talker._is_last_output_token(_info(max_tokens=None), span=1, samples_frame=True) is False
@@ -20,7 +20,6 @@ from vllm_omni.model_extras import (
    get_output_tensor_range,
    get_x_to_text_model_family,
    should_init_extra_args_for_non_diffusion_stages,
-    should_preserve_reference_image_size,
 )


@@ -231,62 +230,18 @@ def test_ltx_extra_registry_declares_official_guidance_params() -> None:
            "audio_rescale_scale",
            "video_stg_blocks",
            "audio_stg_blocks",
-            "sigmas",
-            "stage_1_sigmas",
-            "stage_2_sigmas",
-            "image_crf",
        }
    )

    for pipeline_name in (
        "LTX2Pipeline",
        "LTX2TwoStagePipeline",
-        "LTX2DistilledOneStagePipeline",
-        "LTX2DistilledTwoStagePipeline",
        "LTX2DistilledPipeline",
    ):
        assert get_extra_body_params(pipeline_name) == expected
        assert get_extra_output_params(pipeline_name) == frozenset()


-@pytest.mark.parametrize(
-    ("model_version", "expected"),
-    [("2", False), ("2.3", False), ("2.5", True)],
-)
-def test_ltx_reference_image_size_policy(tmp_path, model_version: str, expected: bool) -> None:
-    (tmp_path / "model_index.json").write_text(
-        '{"_class_name": "LTX2Pipeline", "model_version": "' + model_version + '"}'
-    )
-
-    assert (
-        should_preserve_reference_image_size(
-            None,
-            model=str(tmp_path),
-        )
-        is expected
-    )
-
-
-def test_reference_image_size_policy_threads_revision(monkeypatch: pytest.MonkeyPatch) -> None:
-    captured = {}
-
-    def fake_policy(*, model, revision=None):
-        captured.update(model=model, revision=revision)
-        return True
-
-    monkeypatch.setattr(
-        "vllm_omni.diffusion.models.ltx2.ltx2_components.preserves_reference_image_size",
-        fake_policy,
-    )
-
-    assert should_preserve_reference_image_size(
-        "LTX2Pipeline",
-        model="org/model",
-        revision="pinned-revision",
-    )
-    assert captured == {"model": "org/model", "revision": "pinned-revision"}
-
-
@pytest.mark.core_model
@pytest.mark.cpu
 def test_ming_flash_omni_extra_registry_declares_request_and_response_params() -> None:
@@ -469,13 +424,6 @@ def test_vace_extra_registry_has_no_pipeline_params() -> None:
 def test_unknown_pipeline_has_empty_extra_registry() -> None:
    assert get_extra_body_params("UnknownPipeline") == frozenset()
    assert get_extra_output_params("UnknownPipeline") == frozenset()
-    assert (
-        should_preserve_reference_image_size(
-            "UnknownPipeline",
-            model="org/model",
-        )
-        is False
-    )
    assert should_init_extra_args_for_non_diffusion_stages("UnknownPipeline") is False


@@ -28,8 +28,6 @@ EXCLUDED_MODELS = [
    "WanPipeline",
    "WanVACEPipeline",
    "LTX2TwoStagePipeline",
-    "LTX2DistilledOneStagePipeline",
-    "LTX2DistilledTwoStagePipeline",
    "LTX2DistilledPipeline",
    "LTX2T2VDMD2Pipeline",
    "LTX2I2VDMD2Pipeline",
@@ -159,10 +159,11 @@ def _run_once(model: str, yaml_path: str, *, async_mode: bool) -> list[dict]:
            wav: bytes = b""
            # Batch size 1: submit a single prompt so scheduling is deterministic.
            for step in omni.generate([_text_prompt(question)], _sampling_params(), py_generator=False):
+                out = step.request_output
                if step.final_output_type == "text":
-                    text = step.outputs[0].text  # exact text, no strip -> truly bit-identical
+                    text = out.outputs[0].text  # exact text, no strip -> truly bit-identical
                elif step.final_output_type == "audio":
-                    wav = _audio_bytes(step.outputs[0].multimodal_output["audio"])
+                    wav = _audio_bytes(out.outputs[0].multimodal_output["audio"])
            if text is None:
                raise RuntimeError(f"prompt_id={prompt_id} async={async_mode}: produced no text output")
            if not wav:
@@ -19,10 +19,6 @@ ALLOWED_FILES = {
    "tests/helpers/process.py",
    "vllm_omni/diffusion/distributed/group_coordinator.py",
    "tests/diffusion/attention/test_attention_sp.py",
-    # Measures the wire size of diffusion outputs. The transport pickles them,
-    # and only pickle reproduces the per-view storage duplication under test
-    # (torch.save deduplicates shared storages). No untrusted data is loaded.
-    "tests/diffusion/test_diffusion_ipc.py",
 }

 PICKLE_RE = re.compile(
@@ -15,7 +15,6 @@ from transformers import PretrainedConfig
 from vllm.logger import init_logger
 from vllm.transformers_utils.config import get_config
 from vllm.transformers_utils.repo_utils import get_hf_file_to_dict
-from vllm.transformers_utils.runai_utils import ObjectStorageModel, is_runai_obj_uri

 from vllm_omni.config.endpoint_policy import EndpointRestriction
 from vllm_omni.config.omni_config import VllmOmniConfig
@@ -46,39 +45,6 @@ logger = init_logger(__name__)
 _DEFAULT_PARALLEL_DEGREE = 1


-@functools.cache
-def _materialize_object_storage_configs(model: str) -> str:
-    """Materialize an object-storage model URI's config files locally.
-
-    vLLM's Run:AI streamer keeps ``s3://``/``gs://``/``az://`` URIs opaque until
-    each stage builds its ``ModelConfig``; parent-process resolution (HF config
-    lookup, pipeline/pipeline-key matching) would instead hand the URI to
-    ``huggingface_hub`` helpers, which reject it with ``HFValidationError``.
-    Pull the lightweight files once into vLLM's deterministic
-    ``model_streamer/<hash>`` directory so config reads work here, and so the
-    stage processes' own pull lands in that same directory.
-
-    Returns the input unchanged for non object-storage paths.
-    """
-    if not is_runai_obj_uri(model):
-        return model
-    object_storage_model = ObjectStorageModel(url=model)
-    object_storage_model.pull_files(model, allow_pattern=["*.model", "*.py", "*.json"])
-    logger.info("Materialized object-storage configs for %s at %s", model, object_storage_model.dir)
-    return object_storage_model.dir
-
-
-def _name_match_candidate(model: str) -> str:
-    """Last path component of a model reference, used for name-based matching.
-
-    Object-storage URIs and HF repo ids carry non-model segments (bucket name,
-    organization) that must not participate in substring matching; e.g. a
-    bucket named ``qwen3-tts-models`` holding a ``Qwen3-Omni`` checkpoint must
-    not resolve to the ``qwen3_tts`` pipeline.
-    """
-    return model.rstrip("/").rsplit("/", 1)[-1]
-
-
 def with_trust_remote_code_override(
    overrides: Mapping[str, Any],
    trust_remote_code: bool | None,
@@ -151,7 +117,7 @@ class StageConfigFactory:
        """
        hf_config = None
        try:
-            return get_config(_materialize_object_storage_configs(model), trust_remote_code=trust_remote_code)
+            return get_config(model, trust_remote_code=trust_remote_code)
        except Exception as e:
            logger.debug(f"`get_config` failed with exception {e}; inferred HF config is None")
        return hf_config
@@ -197,12 +163,10 @@ class StageConfigFactory:
        if hf_config is not None:
            return hf_config.model_type

-        config_source = _materialize_object_storage_configs(model)
-
        # Fallback: read config.json directly for custom model types that
        # are not registered with transformers (e.g. qwen3_tts).
        try:
-            config_dict = get_hf_file_to_dict("config.json", config_source, revision=None)
+            config_dict = get_hf_file_to_dict("config.json", model, revision=None)
            if config_dict:
                if "model_type" in config_dict:
                    return config_dict["model_type"]
@@ -219,7 +183,7 @@ class StageConfigFactory:
        # model_index.json with _class_name that maps to a pipeline key via
        # PipelineConfig.diffusers_class_name.
        try:
-            model_index = get_hf_file_to_dict("model_index.json", config_source, revision=None)
+            model_index = get_hf_file_to_dict("model_index.json", model, revision=None)
            if model_index and "_class_name" in model_index:
                class_name = model_index["_class_name"]
                for obj in OMNI_PIPELINES.values():
@@ -239,10 +203,8 @@ class StageConfigFactory:
        # Final fallback: some models (e.g. CosyVoice3) ship an empty
        # config.json and rely on naming conventions. Match the model path
        # basename against registered pipeline keys — longest match wins
-        # so "cosyvoice3" (length 10) beats "cosyvoice" (length 9). Only
-        # the basename is scanned so URI segments such as the bucket name
-        # cannot select an unrelated pipeline.
-        model_lower = _name_match_candidate(model).lower().replace("-", "").replace("_", "")
+        # so "cosyvoice3" (length 10) beats "cosyvoice" (length 9).
+        model_lower = model.lower().replace("-", "").replace("_", "")
        best: str | None = None
        best_len = 0
        for registered_key in OMNI_PIPELINES.keys():
@@ -51,7 +51,6 @@ from vllm_omni.model_executor.models.dots_tts.pipeline import DOTS_TTS_PIPELINE
 from vllm_omni.model_executor.models.dreamzero.pipeline import DREAMZERO_PIPELINE
 from vllm_omni.model_executor.models.dynin_omni.pipeline import DYNIN_OMNI_PIPELINE
 from vllm_omni.model_executor.models.fish_speech.pipeline import FISH_SPEECH_PIPELINE
-from vllm_omni.model_executor.models.gepard.pipeline import GEPARD_PIPELINE
 from vllm_omni.model_executor.models.glm_image.pipeline import GLM_IMAGE_PIPELINE
 from vllm_omni.model_executor.models.glm_tts.pipeline import GLM_TTS_PIPELINE
 from vllm_omni.model_executor.models.gr00t.pipeline import GR00T_N1D7_PIPELINE
@@ -139,7 +138,6 @@ OMNI_PIPELINES: dict[str, PipelineConfig | PipelineResolverFunc] = {
    "lance": LANCE_PIPELINE,
    "dreamzero": DREAMZERO_PIPELINE,
    "Gr00tN1d7": GR00T_N1D7_PIPELINE,
-    "gepard": GEPARD_PIPELINE,
    "glm_image": GLM_IMAGE_PIPELINE,
    "hunyuan_image_3_moe": HUNYUAN_IMAGE3_PIPELINE,
    "hunyuan_image3_ar": HUNYUAN_IMAGE3_AR_PIPELINE,
@@ -1,40 +0,0 @@
-# Gepard-1.0: single-stage native-AR FSQ/NanoCodec TTS (~0.5B).
-# Target 1x L4 24GB. Zero-shot only; enforce_eager until the perf PR.
-#
-# The pipeline pin is required, not cosmetic: the checkpoint's config.json
-# self-identifies as qwen3_5_text, so model_type inference cannot reach the
-# gepard registry entry, and the architectures fallback would match
-# Qwen3_5ForCausalLM against the diffusion pipeline.
-pipeline: gepard
-async_chunk: false
-dtype: bfloat16
-trust_remote_code: true
-
-stages:
-  - stage_id: 0
-    max_num_seqs: 4
-    gpu_memory_utilization: 0.3      # AR LM + NanoCodec fit in ~2 GiB
-    enforce_eager: true
-    enable_prefix_caching: false
-    async_scheduling: true
-    max_num_batched_tokens: 4096
-    max_model_len: 4096
-    devices: "0"
-    # Override architectures so vLLM resolves our talker rather than the native
-    # Qwen3.5 text model. Do NOT override model_type: the backbone's own
-    # model_type check needs it to stay qwen3_5_text.
-    hf_overrides:
-      architectures: ["GepardTalkerForConditionalGeneration"]
-    default_sampling_params:
-      # All 32 heads are sampled in-model, so vLLM's sampler is pinned
-      # deterministic and just argmaxes what forward chose.
-      temperature: 0.0
-      top_p: 1.0
-      top_k: -1
-      # One output token is one frame, so this is the frame budget: 1000 frames
-      # is ~46 s at 21.5 fps. A backstop, not a target -- normal utterances are
-      # 40-420 frames -- but a load-bearing one, because this checkpoint's stop
-      # head does not fire on every batched trajectory.
-      max_tokens: 1000
-      seed: 42
-      detokenize: false
Author	SHA1	Message	Date
Hongsheng Liu	7ec78360bb	Merge branch 'main' into codex-split-attention-offload-guides	2026-08-15 07:00:22 +08:00
Hongsheng Liu	5569abf0a2	Update docs/design/feature/offloader/README.md Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-13 23:37:08 +08:00
Hongsheng Liu	10853fb4f3	Update docs/design/feature/offloader/README.md Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-13 23:37:01 +08:00
Hongsheng Liu	a52c19ea8b	docs: organize offloader feature designs Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-12 11:35:07 +08:00
Hongsheng Liu	8b61499662	docs: rename offloader guides Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-12 09:30:34 +08:00
Hongsheng Liu	7872663bf0	docs: split diffusion feature guides Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-12 08:55:41 +08:00