[Rebase] Rebase to vllm 0.27.0 (#5976 )

Signed-off-by: tzhouam <tzhouam@connect.ust.hk> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
[Test] Add tiny model builder for FluxKontextPipeline (#5823 )
2026-08-11 15:23:46 +00:00 · 2026-08-11 10:59:09 -04:00 · 2026-08-11 14:23:43 +00:00 · 2026-08-11 21:11:39 +08:00 · 2026-08-11 12:46:42 +00:00 · 2026-08-11 17:19:09 +08:00
301 changed files with 23396 additions and 3805 deletions
@@ -8,7 +8,7 @@
 # NPU naming: {queue}_{resource_class}
 #   a2b3_npu_1  — ascend-a2b3, resource_class npu-1 (A2/B3 CI image)
 #   a2b3_npu_4  — ascend-a2b3, resource_class npu-4
-#   a3_npu_2    — ascend-a3, resource_class npu-2 (A3 CI image tag …:$BUILDKITE_COMMIT-a3)
+#   a3_npu_2    — ascend-a3, resource_class npu-2

 x-npu-k8s-plugins: &npu_k8s_plugins
  - kubernetes:
@@ -139,6 +139,11 @@ steps:
        timeout_in_minutes: 50
        commands:
          - pytest -s -v tests/e2e/offline_inference/test_minicpmo_4_5.py tests/e2e/online_serving/test_minicpmo_4_5.py -m 'advanced_model and cuda' --run-level 'advanced_model'
+        # h100_2, not h100_1: advanced_model asserts transcribe every audio
+        # response with Whisper, and tests/helpers/media.py only uses an
+        # accelerator for Whisper when a spare GPU exists. On a single card it
+        # grades on CPU: ~8-10 min per clip vs ~30 s, which alone overruns the
+        # 50-minute job timeout (build 2956).
        mirror_hardwares: h100_2

      - label: "Omni · MiniCPM-o 4.5 Duplex Test"
@@ -148,10 +153,10 @@ steps:
          - tests/e2e/online_serving/test_minicpmo_4_5_duplex.py
          - tests/e2e/online_serving/test_minicpmo_4_5_duplex_expansion.py
          - tests/e2e/online_serving/test_minicpmo_realtime_duplex_drivers.py
-          - tests/e2e/online_serving/minicpmo_realtime_duplex_scenarios.py
+          - 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
-          - tests/helpers/minicpmo_4_5_duplex.py
+          - 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/
@@ -164,7 +169,7 @@ steps:
        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_2
+        mirror_hardwares: h100_1

      - label: "TTS · VoxCPM2 Test"
        source_file_dependencies:
@@ -22,7 +22,7 @@ steps:
        timeout_in_minutes: 50
        commands:
          - pytest -s -v tests/e2e/online_serving/test_minicpmo_4_5_duplex_expansion.py -m "full_model and cuda and H100 and omni" --run-level "full_model"
-        mirror_hardwares: h100_2
+        mirror_hardwares: h100_1

      - label: ":full_moon: Omni · Doc Test with L4"
        timeout_in_minutes: 90
@@ -50,6 +50,19 @@ steps:
            exit $$EXIT
        mirror_hardwares: h100_2

+      - label: ":full_moon: Omni · MiniCPM-o 4.5 · Accuracy Test"
+        timeout_in_minutes: 180
+        commands:
+          - export SEED_TTS_WER_EVAL=1
+          - export SEED_TTS_EVAL_DEVICE=cuda:1
+          - |
+            set +e
+            pytest -s -v tests/e2e/accuracy/minicpmo_4_5/test_minicpmo_4_5.py -m "full_model and cuda" --run-level full_model
+            EXIT=$$?
+            buildkite-agent artifact upload "tests/e2e/accuracy/minicpmo_4_5/results/*.json"
+            exit $$EXIT
+        mirror_hardwares: h100_2
+
      - label: ":full_moon: Omni · Perf Test · No Async Chunk"
        key: nightly-omni-performance-no-async-chunk
        timeout_in_minutes: 300
@@ -102,6 +115,32 @@ steps:
            exit $$EXIT
        mirror_hardwares: h100_3

+      - label: ":full_moon: Omni · MiniCPM-o 4.5 · Perf Test"
+        key: nightly-omni-performance-minicpmo-4-5
+        timeout_in_minutes: 180
+        commands:
+          - export BENCHMARK_DIR=tests/dfx/perf/results
+          - |
+            set +e
+            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_minicpmo_4_5.json
+            EXIT=$$?
+            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
+            exit $$EXIT
+        mirror_hardwares: h100_1
+
+      - label: ":full_moon: Omni · MiniCPM-o 4.5 · Duplex Seed-TTS Perf Test"
+        key: nightly-omni-performance-minicpmo-4-5-duplex-seed-tts
+        timeout_in_minutes: 180
+        commands:
+          - export BENCHMARK_DIR=tests/dfx/perf/results
+          - |
+            set +e
+            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_minicpmo_4_5_duplex_seed_tts.json
+            EXIT=$$?
+            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
+            exit $$EXIT
+        mirror_hardwares: h100_1
+
      - label: ":full_moon: Omni · Multi-Replica Startup Test with 4x H100"
        timeout_in_minutes: 45
        commands:
@@ -404,7 +443,7 @@ steps:
      - label: ":full_moon: Diffusion X2V · LingBot Video MoE Function Test"
        timeout_in_minutes: 90
        commands:
-          - pytest -s -v tests/e2e/online_serving/test_lingbot_video_moe.py -m "full_model and cuda" --run-level "full_model"
+          - pytest -s -v tests/e2e/online_serving/test_lingbot_video_moe.py -k "video_generation" -m "full_model and cuda" --run-level "full_model"
        mirror_hardwares: h100_1

      - label: ":full_moon: Diffusion X2V · Doc Test"
@@ -427,10 +466,10 @@ steps:
          - pytest -s -v tests/e2e/accuracy/test_diffusers_backend_similarity.py -k '2v_matches_diffusers' -m full_model --run-level full_model
        mirror_hardwares: h100_2

-      - label: ":full_moon: Diffusion X2V · MiniMax H3 T2VA Accuracy Test"
+      - label: ":full_moon: Diffusion X2V · MiniMax H3 I2VA/Ref2VA Accuracy Test"
        timeout_in_minutes: 180
        commands:
-          - pytest -s -v tests/e2e/accuracy/minimax_h3/test_minimax_h3_t2va_similarity.py -m full_model --run-level full_model
+          - pytest -s -v tests/e2e/accuracy/minimax_h3/test_minimax_h3_i2va_ref2va_similarity.py -m full_model --run-level full_model
        mirror_hardwares: h100_4

      - label: ":full_moon: Diffusion X2V · Other Accuracy Test"
@@ -128,6 +128,21 @@ steps:
              image: public.ecr.aws/q9t5s3a7/vllm-ci-test-repo:$BUILDKITE_COMMIT
              always-pull: true
              propagate-environment: true
+              # vLLM 0.27 added a pre-flight capacity check in
+              # distributed/device_communicators/shm_broadcast.py. Without this
+              # the container gets Docker's 64 MiB default and every test in
+              # this step dies at engine start with
+              #   RuntimeError: Insufficient space in /dev/shm:
+              #                 240 MiB required, 64 MiB available
+              # (9 failed in 8.7s on build #2949). vLLM 0.26 had no such check
+              # and allocated regardless, so this step passed on main while
+              # being under-provisioned the whole time — 0.27 turned a silent
+              # risk into an explicit error.
+              #
+              # This is the only inline docker plugin block in .buildkite/cuda/;
+              # every other step inherits common/ci_mirror_hardwares.yml, which
+              # already sets 8gb. Matching that rather than inventing a value.
+              shm-size: "8gb"
              environment:
                - "HF_HOME=/fsx/hf_cache"
              volumes:
@@ -172,7 +187,7 @@ steps:
            timeout 20m bash -c "
              pytest -s -v tests/e2e/offline_inference/test_minicpmo_4_5.py -m 'core_model' --run-level 'core_model'
            "
-        mirror_hardwares: h100_2
+        mirror_hardwares: h100_1

      - label: "Omni · MiniCPM-o 4.5 Online Test"
        source_file_dependencies:
@@ -183,7 +198,7 @@ steps:
            timeout 20m bash -c "
              pytest -s -v tests/e2e/online_serving/test_minicpmo_4_5.py -m 'core_model' --run-level 'core_model'
            "
-        mirror_hardwares: h100_2
+        mirror_hardwares: h100_1

      - label: "Omni · MiniCPM-o 4.5 Duplex Test"
        source_file_dependencies:
@@ -192,10 +207,10 @@ steps:
          - tests/e2e/online_serving/test_minicpmo_4_5_duplex.py
          - tests/e2e/online_serving/test_minicpmo_4_5_duplex_expansion.py
          - tests/e2e/online_serving/test_minicpmo_realtime_duplex_drivers.py
-          - tests/e2e/online_serving/minicpmo_realtime_duplex_scenarios.py
+          - 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
-          - tests/helpers/minicpmo_4_5_duplex.py
+          - 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/
@@ -210,7 +225,7 @@ steps:
            timeout 20m bash -c "
              pytest -s -v tests/e2e/online_serving/test_minicpmo_4_5_duplex.py -m 'core_model and cuda' --run-level 'core_model'
            "
-        mirror_hardwares: h100_2
+        mirror_hardwares: h100_1

      - label: "TTS · VoxCPM2 Test"
        source_file_dependencies:
@@ -39,7 +39,7 @@ steps:
            EXIT=$$?
            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
            exit $$EXIT
-      - label: ":robot: Omni · MiniCPM-o 4.5 Test"
+      - label: ":full_moon: Omni · MiniCPM-o 4.5 Test"
        mirror_hardwares: a3_npu_2
        env:
          VLLM_WORKER_MULTIPROC_METHOD: spawn
@@ -47,6 +47,34 @@ steps:
          HF_TOKEN: "${HF_TOKEN}"
        commands:
          - pytest -s -v tests/e2e/offline_inference/test_minicpmo_4_5.py tests/e2e/online_serving/test_minicpmo_4_5.py tests/e2e/online_serving/test_minicpmo_4_5_expansion.py -m 'full_model' --run-level 'full_model'
+      - label: ":full_moon: Omni · MiniCPM-o 4.5 · Duplex Test"
+        timeout_in_minutes: 180
+        mirror_hardwares: a3_npu_2
+        env:
+          VLLM_WORKER_MULTIPROC_METHOD: spawn
+          VLLM_ALLOW_LONG_MAX_MODEL_LEN: "1"
+          HF_TOKEN: "${HF_TOKEN}"
+        commands:
+          - >-
+            pytest -s -v
+            tests/e2e/online_serving/test_minicpmo_4_5_duplex.py
+            tests/e2e/online_serving/test_minicpmo_4_5_duplex_expansion.py
+            -m "npu and A3 and omni"
+            --run-level full_model
+      - label: ":full_moon: Omni · MiniCPM-o 4.5 · Accuracy Test"
+        timeout_in_minutes: 180
+        mirror_hardwares: a3_npu_2
+        env:
+          VLLM_WORKER_MULTIPROC_METHOD: spawn
+          VLLM_ALLOW_LONG_MAX_MODEL_LEN: "1"
+          HF_TOKEN: "${HF_TOKEN}"
+        commands:
+          - |
+            set +e
+            pytest -s -v tests/e2e/accuracy/minicpmo_4_5/test_minicpmo_4_5.py -m "full_model and npu and A3" --run-level full_model
+            EXIT=$$?
+            buildkite-agent artifact upload "tests/e2e/accuracy/minicpmo_4_5/results/*.json"
+            exit $$EXIT
  - group: ":card_index_dividers: TTS Model Test"
    key: nightly-tts-test-group
    depends_on: upload-nightly-pipeline
@@ -3,22 +3,20 @@ steps:
    key: ready-omni-group
    depends_on: upload-ready-pipeline
    steps:
-      - label: ":robot: Omni · MiniCPM-o 4.5 Test"
-        agents:
-          queue: "ascend-a3"
-          resource_class: "npu-2"
-        image: "swr.cn-southwest-2.myhuaweicloud.com/modelfoundry/vllm-omni-ci-npu:${BUILDKITE_COMMIT}-a3"
-        plugins:
-          - kubernetes:
-              podSpecPatch:
-                imagePullSecrets:
-                  - name: swr-secret
+      - label: "Omni · MiniCPM-o 4.5 Test"
+        mirror_hardwares: a3_npu_2
        env:
          VLLM_WORKER_MULTIPROC_METHOD: spawn
          VLLM_ALLOW_LONG_MAX_MODEL_LEN: "1"
-          HF_TOKEN: "${HF_TOKEN}"
        commands:
-          - pytest -s -v tests/e2e/offline_inference/test_minicpmo_4_5.py tests/e2e/online_serving/test_minicpmo_4_5.py -m 'core_model' --run-level 'core_model'
+          - pytest -s -v tests/e2e/offline_inference/test_minicpmo_4_5.py tests/e2e/online_serving/test_minicpmo_4_5.py -m 'core_model and npu and A3 and omni' --run-level 'core_model'
+      - label: "Omni · MiniCPM-o 4.5 Duplex Test"
+        mirror_hardwares: a3_npu_2
+        env:
+          VLLM_WORKER_MULTIPROC_METHOD: spawn
+          VLLM_ALLOW_LONG_MAX_MODEL_LEN: "1"
+        commands:
+          - pytest -s -v tests/e2e/online_serving/test_minicpmo_4_5_duplex.py -m 'core_model and npu and A3 and omni' --run-level 'core_model'

  - group: ":card_index_dividers: TTS Model Test"
    key: ready-tts-group
@@ -1,9 +1,9 @@
-# Claude Skills for vLLM-Omni
+# Repository Skills for vLLM-Omni

-This directory contains Claude Code skills maintained for the `vllm-omni`
-repository. These skills capture repeatable workflows for common contributor
-tasks such as model integration, pull request review, and release note
-generation.
+This directory contains repository-scale skills maintained for `vllm-omni`.
+They capture repeatable workflows for common contributor and maintainer tasks
+such as model integration, CI-aligned testing, performance optimization, and
+pull request review.

 ## Directory Structure

@@ -14,31 +14,41 @@ include:
 - `references/`: focused reference material used by the skill
 - `scripts/`: small helper scripts used by the skill

+## Using the Skills
+
+Coding agents that discover repository skills can invoke the relevant skill by
+name. For other agents, point them to the skill's `SKILL.md` and ask them to
+read it before changing code. Combine a domain skill with `vllm-omni-test` when
+tests or CI coverage change, then use `precheck-pr` for the contributor's final
+self-check.
+
 ## Available Skills

- `add-diffusion-model`: guides integration of a new diffusion model into
-  `vllm-omni`
- `diffusion-perf-opt`: guides diffusion model performance optimization,
-  including profiling traces, parallel strategies, stage timing analysis, and
-  benchmark-driven tuning
- `quantization`: guides quantization method selection, model integration,
-  checkpoint loading, and quality/performance validation for vLLM-Omni
- `add-omni-model`: covers addition of new omni-modality model support
- `add-tts-model`: covers integration of new TTS models and related serving
-  workflows
- `generate-release-note`: helps prepare release notes for repository changes
- `precheck-pr`: self-check a branch before creating a PR — validates PR title
-  format, catches dead code, verifies accuracy/perf claims, and confirms merge
-  readiness
- `vllm-omni-test`: guides generation and execution of CI-aligned tests (L1–L4),
-  pytest marker selection (`core_model` / `advanced_model` / `full_model`,
-  `omni` / `tts` / `diffusion`), Buildkite wiring (`test-ready.yml`,
-  `test-merge.yml`, `test-nightly.yml`, `test-weekly.yml`), and copy-paste
-  local plus CI-like `pytest` commands; see `references/test-routing.md` for
-  level-to-command mapping
- `review-pr`: provides a structured workflow for reviewing pull requests
- `vllm-omni-npu-model-runner-upgrade`: upgrades NPU model runners to align with the
-  latest vllm-ascend NPUModelRunner
+- [`add-diffusion-model`](add-diffusion-model/SKILL.md): guides integration of
+  a new diffusion model into `vllm-omni`
+- [`add-tts-model`](add-tts-model/SKILL.md): covers integration of new TTS
+  models and related serving workflows
+- [`diffusion-perf-opt`](diffusion-perf-opt/SKILL.md): guides diffusion model
+  performance optimization, including profiling traces, parallel strategies,
+  stage timing analysis, and benchmark-driven tuning
+- [`precheck-pr`](precheck-pr/SKILL.md): self-checks a branch before creating a
+  PR by validating title format, dead code, accuracy and performance claims,
+  and merge readiness
+- [`quantization`](quantization/SKILL.md): guides quantization method selection,
+  model integration, checkpoint loading, and quality/performance validation
+  for vLLM-Omni
+- [`review-pr`](review-pr/SKILL.md): provides a frozen-snapshot, contract-aware
+  workflow for maintainers and reviewers; contributors should use
+  `precheck-pr` for self-review
+- [`vllm-omni-npu-model-runner-upgrade`](vllm-omni-npu-upgrade/SKILL.md):
+  upgrades NPU model runners to align with the latest vllm-ascend
+  `NPUModelRunner`
+- [`vllm-omni-test`](vllm-omni-test/SKILL.md): guides generation and execution
+  of CI-aligned tests (L1–L4), pytest marker selection (`core_model` /
+  `advanced_model` / `full_model`, `omni` / `tts` / `diffusion`), Buildkite
+  wiring (`test-ready.yml`, `test-merge.yml`, `test-nightly.yml`,
+  `test-weekly.yml`), and copy-paste local plus CI-like `pytest` commands; see
+  `references/test-routing.md` for level-to-command mapping

 ## Maintenance Guidelines

@@ -94,7 +94,6 @@ If docs and live code disagree, verify the code/tests and report the drift.
 | --- | --- |
 | [model-addition-checklist.md](references/checks/model-addition-checklist.md) | A model, architecture, loader, processor, registry, or stage config is added. |
 | [perf-verification.md](references/checks/perf-verification.md) | The PR makes a latency, throughput, memory, or quality claim. |
-| [config-provenance.md](references/checks/config-provenance.md) | Changed or newly load-bearing behavior reads config, CLI, environment, legacy, or default values. |
 | [test-quality-evaluation.md](references/checks/test-quality-evaluation.md) | Tests change, are absent for risky code, or may not exercise production behavior. |
 | [tests-docs-checklist.md](references/checks/tests-docs-checklist.md) | Coverage, CI markers, examples, user docs, or PR evidence need review. |
 | [verification.md](references/checks/verification.md) | Hardware, a server, or a runnable affected path is available for active verification. |
@@ -153,19 +152,13 @@ contract they protect or use only the applicable evidence checks.
 Apply every category in [general-checks.md](references/process/general-checks.md) before
 lower-priority comments.

-For each changed or newly load-bearing value or behavior, trace:
+For each changed value or behavior, trace:

 ```text
 public ingress -> validation/defaulting -> producer -> transformations
  -> stage/worker/connector boundary -> final consumer -> terminal cleanup
 ```

-Treat unchanged context as in scope when the diff makes it authoritative for a
-new consumer. When config, CLI, environment, legacy, or default sources can
-control that value, apply
-[config-provenance.md](references/checks/config-provenance.md) and prove the
-documented source at the final sink with distinct sentinel values.
-
 Cover every applicable offline/online, streaming/non-streaming, sync/async,
 feature-on/off, topology, and compatibility path. Search bounded callers and
 sibling implementations rather than assuming the changed hunk is the only path.
@@ -1,49 +0,0 @@
-# Configuration Provenance
-
-Read this reference when changed or newly load-bearing behavior reads config,
-CLI, environment, legacy, alias, or default values. Always use it for endpoint,
-host, port, address, connector metadata, and `*_extra_config` consumers, even
-when the schema or documentation is unchanged.
-
-## Trace the source of truth
-
-For each semantic value, record one compact ledger:
-
-```text
-documented input -> parsed field -> competing sources and precedence
-  -> runtime projection/transport -> final consumer
-```
-
- Search the exact field, environment variable, default literal, and aliases in
-  code, tests, docs, examples, and deployment files.
- Treat unchanged producers and context lines as review scope when the diff
-  starts relying on them or makes their output authoritative.
- Resolve the intended precedence from active branch-local docs, enforced
-  code/tests, and compatibility policy. Report a documented input that is
-  silently ignored or reinterpreted at a live sink.
- Check every supported construction path and feature-off behavior without
-  inventing parity for unsupported paths.
-
-## Require discriminating evidence
-
-Give every competing source a distinct value so the test reveals which source
-wins. Prefer non-default sentinels; for example:
-
-```text
-documented config port = 25301
-environment/default port = 8998
-expected final endpoint = http://10.0.0.8:25301
-```
-
-Assert the final URL, connection target, runtime argument, or other live sink,
-not only an intermediate object. A test that changes the environment value to
-the expected configured value proves formatting, not configuration propagation.
-For a bug fix, show the discriminating test fails on the frozen base and passes
-on the reviewed head when the environment permits execution.
-
-## Finding bar
-
-A supported non-default value reaching the wrong endpoint, device, topology, or
-runtime behavior is a correctness or compatibility finding. A missing
-discriminating test alone is a validation gap unless repository policy requires
-that regression coverage for the change.
@@ -16,9 +16,6 @@ were reverted or broken. Check:
  instead of the changed behavior being mocked away;
 - fakes preserve relevant types, MRO, shapes, devices, async behavior, and
  lifecycle transitions;
- configuration tests give competing config, CLI, environment, legacy, and
-  default sources distinct non-default values and assert the final consumer;
-  setting only the source read by the implementation does not prove precedence;
 - seeds, ordering, timing, synchronization, external services, and numeric
  tolerances are deterministic or explicitly controlled;
 - normal, invalid, boundary, feature-off, failure/cancellation, and regression
@@ -15,9 +15,6 @@ and [design documents](https://docs.vllm.ai/projects/vllm-omni/en/latest/design/
 - Trace each changed value through public ingress, validation/defaulting,
  producer, transformations, process or device boundaries, final consumer, and
  terminal cleanup.
- Treat unchanged code as in scope when the diff makes its value newly
-  load-bearing for a changed consumer; validate that value's origin and contract
-  rather than reviewing only added lines.
 - Check every applicable sync/async, offline/online, streaming/non-streaming,
  feature-off, topology, and compatibility path without demanding unsupported
  modes merely for symmetry.
@@ -126,9 +123,7 @@ distinct live caller, invariant, or compatibility need.

 ## Finding bar

-Anchor each finding to the exact `path:line` that causes the behavior; it may be
-unchanged when the diff makes it newly load-bearing, in which case name the
-changed call path that brings it into scope. Include the trigger, impact, and
-smallest fix direction. Treat pending CI, missing hardware, or unsupported
-measurements as validation gaps unless the repository contract makes that
-evidence a merge requirement.
+Anchor each finding to a changed `path:line`; name the trigger or call path,
+current behavior, impact, and smallest fix direction. Treat pending CI, missing
+hardware, or unsupported measurements as validation gaps unless the repository
+contract makes that evidence a merge requirement.
@@ -166,10 +166,6 @@ index patch, worktree patch, and untracked-content manifest. Any mismatch makes
 the review stale; discard the affected evidence and restart. Prefer one
 root-cause comment to several symptom comments.

-When the root cause is an unchanged line made newly load-bearing by the diff,
-cite it in the consolidated review body instead of moving the comment to an
-unrelated changed line merely to make it inline.
-
 Local presentation is the default. Only explicit authorization permits GitHub
 posting. When authorized, publish one consolidated final review after
 validation; do not post preliminary or incremental comments. Submit a review
@@ -50,7 +50,6 @@ compatibility changes.
 | --- | --- | --- |
 | New or expanded model, loader, processor, registry, or stage config | [model-addition-checklist.md](../checks/model-addition-checklist.md) | [`add-tts-model`](../../../add-tts-model/SKILL.md) or [`add-diffusion-model`](../../../add-diffusion-model/SKILL.md) |
 | Latency, throughput, memory, scaling, precision, or quality claim | [perf-verification.md](../checks/perf-verification.md) | [`diffusion-perf-opt`](../../../diffusion-perf-opt/SKILL.md) for diffusion |
-| Config/CLI/environment/default precedence; endpoint, host, port, address, connector metadata, or `*_extra_config` | [config-provenance.md](../checks/config-provenance.md) | None |
 | Tests changed, absent for risky behavior, or test-only | [test-quality-evaluation.md](../checks/test-quality-evaluation.md) | [`vllm-omni-test`](../../../vllm-omni-test/SKILL.md) |
 | CI, examples, docs, public behavior, or contributor evidence | [tests-docs-checklist.md](../checks/tests-docs-checklist.md) | None |
 | Suitable hardware/server and runnable affected path | [verification.md](../checks/verification.md) | None |
@@ -1,6 +1,11 @@
 # CODEOWNERS is used to automatically assign PR reviewers
 # Format: path pattern @username or @team
 # More info: https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/about-code-owners
+#
+# The vllm_omni/** source-tree rules below are derived from the `owners` and
+# `primary_code_paths` front matter in docs/design/module/*.md. When updating
+# ownership, update the design doc first, then regenerate the matching rule here.
+# Last-match-wins: more specific rules must come AFTER broader ones.

 # Default owners (fallback reviewers for all files)
 * @Gaohan123 @hsliuustc0106 @ywang96
@@ -21,25 +26,10 @@
 # CI
 /.buildkite/ @yenuo26 @congw729 @NickCao

-# Engine and model_executor
-/vllm_omni/core/ @tzhouam @yinpeiqi @fake0fan @Sy0307
-/vllm_omni/engine/ @tzhouam @yinpeiqi @fake0fan @Sy0307
-/vllm_omni/entrypoints/ @tzhouam @yinpeiqi @fake0fan @alex-jw-brooks @NickCao
-/vllm_omni/model_executor/ @tzhouam @gcanlin
-/vllm_omni/worker/ @tzhouam @gcanlin
+# ComfyUI / external apps
+/apps/ @fhfuih @alex-jw-brooks

-# Diffusion
-/vllm_omni/diffusion/ @Isotr0py @princepride @SamitHuang @wtomin @ZJY0516 @RuixiangMa @david6666666 @xuechendi
-/vllm_omni/lora/ @princepride @SamitHuang @wtomin @ZJY0516 @RuixiangMa @yuanheng-zhao
-
-# Diffusion CPU offloading
-/vllm_omni/diffusion/offloader/ @Isotr0py @princepride @SamitHuang @wtomin @ZJY0516 @RuixiangMa @yuanheng-zhao @xuechendi
-
-# Omni and TTS support
-/vllm_omni/transformers_utils/ @tzhouam @yuanheng-zhao
-/vllm_omni/model_executor/models/ @linyueqian @ZeldaHuang @princepride @yuanheng-zhao @amy-why-3459 @Bounty-hunter @Sy0307 @gcanlin @NickCao
-
-# Benchmarks
+# Benchmarks (out-of-tree)
 /benchmarks/accuracy/ @david6666666
 /benchmarks/build_dataset/ @linyueqian @yuanheng-zhao
 /benchmarks/diffusion/ @Isotr0py @princepride @SamitHuang @wtomin @ZJY0516 @RuixiangMa @xuechendi
@@ -47,20 +37,112 @@
 /benchmarks/glm_image/ @princepride @JaredforReal
 /benchmarks/tts/ @linyueqian @ZeldaHuang @princepride @Sy0307 @gcanlin

-# Usability and Observability
-/vllm_omni/deploy/ @tzhouam @lishunyang12
-/vllm_omni/metrics/ @lishunyang12
-/vllm_omni/profiler/ @gcanlin
+# =====================================================================
+# Module design ownership (docs/design/module/*.md)
+# Each section header cites the source design doc.
+# =====================================================================

-# ComfyUI
-/apps/ @fhfuih @alex-jw-brooks
+# Autoregressive runtime — docs/design/module/ar_runtime.md
+/vllm_omni/core/ @tzhouam @yinpeiqi @fake0fan @Sy0307 @Gaohan123
+/vllm_omni/worker/ @tzhouam @yinpeiqi @fake0fan @Sy0307 @Gaohan123

-# Large Scale Serving
+# Engine orchestration & stage runtime — docs/design/module/engine_orchestration.md, stage_runtime.md
+/vllm_omni/engine/ @tzhouam @fake0fan
+# File-level overrides: input/output modality contracts own these engine files
+# — docs/design/module/input_output_modality_contracts.md
+/vllm_omni/engine/messages.py @Sy0307 @amy-why-3459 @Gaohan123 @alex-jw-brooks
+/vllm_omni/engine/serialization.py @Sy0307 @amy-why-3459 @Gaohan123 @alex-jw-brooks
+/vllm_omni/engine/mm_outputs.py @Sy0307 @amy-why-3459 @Gaohan123 @alex-jw-brooks
+/vllm_omni/engine/output_modality.py @Sy0307 @amy-why-3459 @Gaohan123 @alex-jw-brooks
+/vllm_omni/engine/output_processor.py @Sy0307 @amy-why-3459 @Gaohan123 @alex-jw-brooks
+
+# Input/output modality contracts — docs/design/module/input_output_modality_contracts.md
+/vllm_omni/inputs/ @Sy0307 @amy-why-3459 @Gaohan123 @alex-jw-brooks
+/vllm_omni/outputs/ @Sy0307 @amy-why-3459 @Gaohan123 @alex-jw-brooks
+/vllm_omni/request.py @Sy0307 @amy-why-3459 @Gaohan123 @alex-jw-brooks
+/vllm_omni/data_entry_keys.py @Sy0307 @amy-why-3459 @Gaohan123 @alex-jw-brooks
+
+# Entrypoints and serving boundaries — docs/design/module/entrypoints.md
+/vllm_omni/entrypoints/ @alex-jw-brooks @linyueqian @NickCao
+# File-level overrides within entrypoints/
+/vllm_omni/entrypoints/openai/errors.py @alex-jw-brooks @NickCao
+/vllm_omni/entrypoints/cli/benchmark/ @alex-jw-brooks @Bounty-hunter
+
+# Error contracts — docs/design/module/error_contracts.md
+/vllm_omni/errors.py @alex-jw-brooks @NickCao
+
+# OmniConnector (inter-stage transport) — docs/design/module/omni_connector.md
 /vllm_omni/distributed/ @princepride @yuanheng-zhao @xuechendi @natureofnature @fake0fan

-# Quantization and configuration
+# Model integration — docs/design/module/model_integration.md
+/vllm_omni/model_executor/ @tzhouam @gcanlin
+/vllm_omni/model_executor/models/ @tzhouam @gcanlin @Sy0307 @amy-why-3459 @linyueqian
+/vllm_omni/model_extras/ @tzhouam @gcanlin @princepride
+/vllm_omni/plugins/ @gcanlin @tjtanaa @xuechendi
+# Config override within model_executor — docs/design/module/vllm_omni_config.md
+/vllm_omni/model_executor/stage_configs/ @lishunyang12 @alex-jw-brooks
+
+# Execution platforms (hardware backends) — docs/design/module/execution_platforms.md
+/vllm_omni/platforms/ @gcanlin @tjtanaa @xuechendi
+/vllm_omni/attention/ @gcanlin @tjtanaa @xuechendi @david6666666 @lishunyang12
+
+# Configuration & deploy — docs/design/module/vllm_omni_config.md
 /vllm_omni/config/ @lishunyang12 @alex-jw-brooks
+/vllm_omni/deploy/ @lishunyang12 @alex-jw-brooks
+
+# Quantization — docs/design/module/quantization.md
 /vllm_omni/quantization/ @david6666666 @Isotr0py @lishunyang12

-# Hardware plugins and accelerators
-/vllm_omni/platforms/ @gcanlin @tjtanaa @xuechendi
+# Observability — docs/design/module/observability.md
+/vllm_omni/metrics/ @lishunyang12 @vraiti
+
+# Profiling — docs/design/module/profiling.md
+/vllm_omni/profiler/ @gcanlin
+
+# Benchmarking (in-tree) — docs/design/module/benchmarking.md
+/vllm_omni/benchmarks/ @alex-jw-brooks @Bounty-hunter
+
+# LoRA (no dedicated design doc; ownership intentionally narrowed)
+/vllm_omni/lora/ @SamitHuang @wtomin
+
+# Transformers utils (related path only in design docs; preserves prior ownership)
+/vllm_omni/transformers_utils/ @tzhouam @yuanheng-zhao
+
+# =====================================================================
+# Diffusion module family — docs/design/module/diffusion/
+# Umbrella rule first; sub-module rules after it win via last-match.
+# =====================================================================
+
+# Diffusion umbrella (top-level files: data.py, forward_context.py, etc.)
+# — docs/design/module/diffusion/index.md
+/vllm_omni/diffusion/ @Isotr0py @princepride @SamitHuang @wtomin @ZJY0516 @RuixiangMa @david6666666 @xuechendi
+
+# Cache management — docs/design/module/cache_management.md
+/vllm_omni/diffusion/cache/ @Isotr0py @princepride @SamitHuang
+
+# Diffusion quantization — docs/design/module/quantization.md
+/vllm_omni/diffusion/quantization/ @david6666666 @Isotr0py @lishunyang12
+
+# Diffusion profiling — docs/design/module/profiling.md
+/vllm_omni/diffusion/profiler/ @gcanlin
+
+# Diffusion runtime (scheduler, executor, worker)
+# — docs/design/module/diffusion/diffusion_runtime.md, continuous_batching.md
+/vllm_omni/diffusion/sched/ @Isotr0py @princepride @SamitHuang @fhfuih
+/vllm_omni/diffusion/executor/ @Isotr0py @princepride @SamitHuang @fhfuih
+/vllm_omni/diffusion/worker/ @Isotr0py @princepride @SamitHuang @fhfuih
+
+# Diffusion model integration (pipelines, registration, checkpoint loading)
+# — docs/design/module/diffusion/diffusion_model_integration.md
+/vllm_omni/diffusion/models/ @fhfuih @Bounty-hunter @wtomin
+/vllm_omni/diffusion/model_loader/ @fhfuih @Bounty-hunter @wtomin @RuixiangMa
+/vllm_omni/diffusion/lora/ @fhfuih @Bounty-hunter @wtomin @RuixiangMa
+
+# Diffusion parallelism (rank topology, TP/SP/PP/VAE patch)
+# — docs/design/module/diffusion/parallelism.md
+/vllm_omni/diffusion/distributed/ @princepride @yuanheng-zhao @xuechendi @Bounty-hunter
+/vllm_omni/diffusion/attention/parallel/ @princepride @yuanheng-zhao @xuechendi @Bounty-hunter
+
+# Diffusion offloader (CPU offloading, DLO)
+# — docs/design/module/diffusion/offloader.md
+/vllm_omni/diffusion/offloader/ @yuanheng-zhao @david6666666 @lishunyang12
@@ -66,7 +66,10 @@ repos:
      entry: python tools/pre_commit/check_tts_adapter.py
      language: python
      pass_filenames: false
-      files: ^vllm_omni/entrypoints/openai/(serving_speech\.py|tts_adapters/__init__\.py)$
+      files: ^(vllm_omni/entrypoints/openai/(serving_speech\.py|tts_adapters/__init__\.py)|tools/pre_commit/check_tts_adapter\.py)$
+      # The budget-is-now-slack reminder is printed on a PASSING run, and
+      # pre-commit hides output from passing hooks unless the hook is verbose.
+      verbose: true

    # Keep `suggestion` last
    - id: suggestion
@@ -1,7 +1,24 @@
-# Target vllm version: v0.26.0 (releases/v0.26.0 branch).
+# Target vllm version: v0.27.0 (releases/v0.27.0 branch).
 # Keep VLLM_BASE_TAG aligned with the target vLLM release on every rebase.
+#
+# Release-boundary form: there is no precompiled-wheel pin here, because the
+# base image already ships exactly the target vLLM. Everything the runtime
+# needs — torch, flashinfer, nvJitLink, numpy — comes from that image as one
+# self-consistent set.
+#
+# This block reappears only when the rebase targets a vLLM commit that is NOT a
+# released tag. During the 0.27 rebase, targeting v0.27.0rc1 on top of the
+# v0.26.0 base image required reinstalling vLLM from wheels.vllm.ai and then
+# hand-repairing the fallout: pinning torch ahead of flashinfer (installing it
+# after bound flashinfer's TVM FFI runtime to the old torch and aborted engine
+# core init), registering libnvJitLink.so.13 with ldconfig, and re-pinning numpy
+# under numba's ceiling. All of it was compensation for a base image and a vLLM
+# that disagreed. At a real release boundary they agree, so none of it is
+# needed. If you pin an untagged commit again, restore the reinstall block —
+# agent/tasks/20_update_wheel_url.sh fails the run rather than letting the
+# Dockerfile silently ship the wrong vLLM.
 ARG VLLM_BASE_IMAGE=vllm/vllm-openai
-ARG VLLM_BASE_TAG=v0.26.0
+ARG VLLM_BASE_TAG=v0.27.0

 # We use registry-based layer caching for the dependencies,
 # where the tag is determined based on the hash of the contents
@@ -45,6 +45,7 @@ RUN source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
      "hyperpyyaml>=1.2.0" \
      "onnx>=1.16.0" \
      "librosa>=0.10.0" \
+      "datasets>=3.0.0" \
      --index-url http://cache-service.nginx-pypi-cache.svc.cluster.local/pypi/simple \
      --trusted-host cache-service.nginx-pypi-cache.svc.cluster.local \
      --default-timeout=60 && \
@@ -46,6 +46,7 @@ RUN source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
      "onnx>=1.16.0" \
      "librosa>=0.10.0" \
      "torchmetrics>=1.9.0" \
+      "datasets>=3.0.0" \
      --index-url http://cache-service.nginx-pypi-cache.svc.cluster.local/pypi/simple \
      --trusted-host cache-service.nginx-pypi-cache.svc.cluster.local \
      --default-timeout=60 && \
@@ -16,31 +16,21 @@ nav:
  - Examples:
    - examples/README.md
    - Offline Inference:
-      - GLM-Image Multistage End-to-End Inference: user_guide/examples/offline_inference/glm_image.md
-      - HunyuanImage-3.0 Image-to-Text Inference: user_guide/examples/offline_inference/hunyuan_image3.md
      - Image-To-Image: user_guide/examples/offline_inference/image_to_image.md
      - Image-To-Video: user_guide/examples/offline_inference/image_to_video.md
-      - InternVLA-A1: user_guide/examples/offline_inference/internvla_a1.md
-      - MiMo-Audio: user_guide/examples/offline_inference/mimo_audio.md
-      - Qwen2.5-Omni: user_guide/examples/offline_inference/qwen2_5_omni.md
-      - Qwen3-Omni: user_guide/examples/offline_inference/qwen3_omni.md
+      - Speech-To-Video: user_guide/examples/offline_inference/speech_to_video.md
      - Text-To-Audio: user_guide/examples/offline_inference/text_to_audio.md
      - Text-To-Image: user_guide/examples/offline_inference/text_to_image.md
      - Text-To-Speech: user_guide/examples/offline_inference/text_to_speech.md
      - Text-To-Video: user_guide/examples/offline_inference/text_to_video.md
+      - X-To-Text: user_guide/examples/offline_inference/x_to_text.md
      - X-To-Video-Audio: user_guide/examples/offline_inference/x_to_video_audio.md
    - Online Serving:
-      - BAGEL-7B-MoT: user_guide/examples/online_serving/bagel.md
-      - vLLM-Omni Helm Chart: user_guide/examples/online_serving/chart-helm.md
-      - Diffusers Backend Adapter: user_guide/examples/online_serving/diffusers_pipeline_adapter.md
-      - GLM-Image Online Serving: user_guide/examples/online_serving/glm_image.md
      - Image-To-Image: user_guide/examples/online_serving/image_to_image.md
      - Image-To-Video: user_guide/examples/online_serving/image_to_video.md
-      - MiMo-Audio: user_guide/examples/online_serving/mimo_audio.md
-      - Qwen2.5-Omni: user_guide/examples/online_serving/qwen2_5_omni.md
-      - Qwen3-Omni: user_guide/examples/online_serving/qwen3_omni.md
-      - Text-To-Speech: user_guide/examples/online_serving/text_to_speech.md
+      - Speech-To-Video: user_guide/examples/online_serving/speech_to_video.md
      - Text-To-Image: user_guide/examples/online_serving/text_to_image.md
+      - Text-To-Speech: user_guide/examples/online_serving/text_to_speech.md
      - Text-To-Video: user_guide/examples/online_serving/text_to_video.md
  - General:
    - usage/*
@@ -76,6 +76,47 @@ Please refer to the [test instructions](./ci/test_execution_guide.md) for detail
 !!! warning
    Currently, not all unit tests pass when run on CPU platforms. If you don't have access to a GPU platform to run unit tests locally, rely on the continuous integration system to run the tests for now.

+### Using repository skills with coding agents
+
+vLLM-Omni maintains [repository-scale skills](https://github.com/vllm-project/vllm-omni/tree/main/.claude/skills) for common coding, testing, performance, and review workflows. These skills capture repository-specific structure, validation requirements, and evidence standards that a general coding agent may not know.
+
+If your coding agent discovers repository skills automatically, ask it to use the relevant skill by name. Otherwise, point it to the linked `SKILL.md` and ask it to read the file before changing code. Skills can be combined: use a domain skill for the implementation, `vllm-omni-test` for test and CI decisions, and `precheck-pr` before opening the pull request.
+
+| Task | Repository skill |
+| --- | --- |
+| Add or extend a diffusion model | [`add-diffusion-model`](https://github.com/vllm-project/vllm-omni/blob/main/.claude/skills/add-diffusion-model/SKILL.md) |
+| Add or extend a text-to-speech model | [`add-tts-model`](https://github.com/vllm-project/vllm-omni/blob/main/.claude/skills/add-tts-model/SKILL.md) |
+| Add, select, or debug quantization | [`quantization`](https://github.com/vllm-project/vllm-omni/blob/main/.claude/skills/quantization/SKILL.md) |
+| Diagnose or optimize diffusion performance | [`diffusion-perf-opt`](https://github.com/vllm-project/vllm-omni/blob/main/.claude/skills/diffusion-perf-opt/SKILL.md) |
+| Upgrade an NPU model runner | [`vllm-omni-npu-model-runner-upgrade`](https://github.com/vllm-project/vllm-omni/blob/main/.claude/skills/vllm-omni-npu-upgrade/SKILL.md) |
+| Add a regression test, choose L1-L4 coverage, or wire Buildkite | [`vllm-omni-test`](https://github.com/vllm-project/vllm-omni/blob/main/.claude/skills/vllm-omni-test/SKILL.md) |
+| Self-check a branch before opening a pull request | [`precheck-pr`](https://github.com/vllm-project/vllm-omni/blob/main/.claude/skills/precheck-pr/SKILL.md) |
+| Perform a maintainer or reviewer review of an existing pull request or local branch | [`review-pr`](https://github.com/vllm-project/vllm-omni/blob/main/.claude/skills/review-pr/SKILL.md) |
+
+Use the following workflow when collaborating with a coding agent:
+
+1. **Ground the task.** Provide the issue, RFC, or design document that defines the expected behavior. Ask the agent to state its assumptions and identify the affected module or feature contracts before editing code.
+2. **Select the narrowest applicable skills.** Start with the domain skill that owns the change. Add `vllm-omni-test` whenever production behavior or tests change. For a bug fix, ask for the smallest reproducer and a regression test tied to the original symptom.
+3. **Require executable evidence.** Ask the agent to run the narrowest relevant checks and report the exact commands and results. It must identify tests it could not run because of unavailable hardware, model weights, credentials, or dependencies; an unrun check is not a pass.
+4. **Inspect the resulting diff.** Confirm that the implementation stays within the issue or RFC, does not overwrite unrelated work, and includes the required tests, documentation, and benchmark or accuracy evidence.
+5. **Run the pre-submit workflow.** Use `precheck-pr`, then run the applicable local tests and pre-commit hooks before opening the pull request.
+
+For example:
+
+```text
+Use add-diffusion-model and vllm-omni-test to implement the linked issue.
+Read the relevant design documents first, keep the change within the issue,
+and report every test command, result, and hardware validation gap.
+```
+
+```text
+Use vllm-omni-test to turn this bug reproducer into the smallest deterministic
+regression test. Prefer L1 CPU coverage; if the failure requires real weights
+or serving, explain the required L2/L3 environment and provide the exact command.
+```
+
+Repository skills guide the agent, but they do not replace contributor judgment, the accepted issue or RFC, required test evidence, or maintainer review. Review all generated changes before submitting them.
+
 ## Issues

 If you encounter a bug or have a feature request, please search existing issues first to see if it has already been reported. If not, please file a new issue, providing as much relevant information as possible.
@@ -7,6 +7,7 @@ owners:
  - "@yinpeiqi"
  - "@fake0fan"
  - "@Sy0307"
+  - "@Gaohan123"
 primary_code_paths:
  - vllm_omni/core/**
  - vllm_omni/worker/**
@@ -4,6 +4,7 @@ kind: module
 status: draft
 owners:
  - "@alex-jw-brooks"
+  - "@Bounty-hunter"
 primary_code_paths:
  - vllm_omni/benchmarks/**
  - vllm_omni/entrypoints/cli/benchmark/**
@@ -5,6 +5,7 @@ status: draft
 owners:
  - "@Isotr0py"
  - "@princepride"
+  - "@SamitHuang"
 primary_code_paths:
  - vllm_omni/diffusion/cache/**
 related_code_paths:
@@ -3,9 +3,10 @@ title: Diffusion Model Integration
 kind: module
 status: draft
 owners:
-  - "@Isotr0py"
+  - "@fhfuih"
+  - "@Bounty-hunter"
  - "@wtomin"
-  - "@ZJY0516"
+  - "@RuixiangMa"
 primary_code_paths:
  - vllm_omni/diffusion/models/**
  - vllm_omni/diffusion/model_loader/**
@@ -6,6 +6,7 @@ owners:
  - "@Isotr0py"
  - "@princepride"
  - "@SamitHuang"
+  - "@fhfuih"
 primary_code_paths:
  - vllm_omni/diffusion/sched/**
  - vllm_omni/diffusion/executor/**
@@ -4,7 +4,8 @@ kind: module
 status: draft
 owners:
  - "@yuanheng-zhao"
-  - "@Isotr0py"
+  - "@david6666666"
+  - "@lishunyang12"
 primary_code_paths:
  - vllm_omni/diffusion/offloader/**
 related_code_paths:
@@ -6,6 +6,7 @@ owners:
  - "@princepride"
  - "@yuanheng-zhao"
  - "@xuechendi"
+  - "@Bounty-hunter"
 primary_code_paths:
  - vllm_omni/diffusion/distributed/**
  - vllm_omni/diffusion/attention/parallel/**
@@ -6,6 +6,7 @@ architecture_state: current-plus-in-flight
 owners:
  - "@alex-jw-brooks"
  - "@linyueqian"
+  - "@NickCao"
 document_stewards:
  - "@hsliuustc0106"
  - "@Gaohan123"
@@ -8,8 +8,15 @@ owners:
  - "@xuechendi"
 primary_code_paths:
  - vllm_omni/platforms/**
-related_code_paths:
  - vllm_omni/attention/**
+primary_code_path_owners:
+  # Additional owners scoped to attention kernels (quantized/custom attention)
+  - paths:
+      - vllm_omni/attention/**
+    owners:
+      - "@david6666666"
+      - "@lishunyang12"
+related_code_paths: []
 depends_on:
  - ar_runtime.md
  - diffusion/index.md
@@ -6,6 +6,8 @@ architecture_state: current
 owners:
  - "@Sy0307"
  - "@amy-why-3459"
+  - "@Gaohan123"
+  - "@alex-jw-brooks"
 document_stewards:
  - "@hsliuustc0106"
  - "@Gaohan123"
@@ -9,6 +9,14 @@ primary_code_paths:
  - vllm_omni/model_executor/**
  - vllm_omni/model_extras/**
  - vllm_omni/plugins/**
+primary_code_path_owners:
+  # Additional owners scoped to a sub-path of model_executor/**
+  - paths:
+      - vllm_omni/model_executor/models/**
+    owners:
+      - "@Sy0307"
+      - "@amy-why-3459"
+      - "@linyueqian"
 related_code_paths:
  - vllm_omni/transformers_utils/**
  - vllm_omni/tokenizers/**
@@ -4,6 +4,7 @@ kind: module
 status: draft
 owners:
  - "@lishunyang12"
+  - "@vraiti"
 primary_code_paths:
  - vllm_omni/metrics/**
 related_code_paths:
@@ -123,6 +123,7 @@ def extract_omni_serve_subparser_init():
                        "TrackingNamespace": None,  # type annotation only
                        "make_arg_parser": lambda parser: parser,  # no-op for doc
                        "_ensure_vllm_platform": lambda: None,  # no-op for doc
+                        "_nonneg_finite_float": float,
                        "VLLM_SUBCMD_PARSER_EPILOG": "",
                        "logger": logger,
                        "DummySubparsers": DummySubparsers,
@@ -24,6 +24,23 @@ SERVING_MODE_TITLES = {
    "online_serving": "Online serving",
 }

+# Keep the Examples navigation focused on reusable modality/task entry points.
+# Model-specific scripts remain available under ``examples/`` and are surfaced
+# from these shared pages instead of getting one page per model.
+GENERAL_EXAMPLE_SLUGS = frozenset(
+    {
+        "image_to_image",
+        "image_to_video",
+        "speech_to_video",
+        "text_to_audio",
+        "text_to_image",
+        "text_to_speech",
+        "text_to_video",
+        "x_to_text",
+        "x_to_video_audio",
+    }
+)
+

 def load_model_display_names() -> dict[str, str]:
    try:
@@ -46,6 +63,13 @@ def load_model_display_names() -> dict[str, str]:
 MODEL_DISPLAY_NAMES = load_model_display_names()


+def is_general_example(example: "Example") -> bool:
+    """Return whether an example should get a generated serving page."""
+    if example.category not in SERVING_MODE_TITLES:
+        return True
+    return example.path.stem in GENERAL_EXAMPLE_SLUGS
+
+
 def fix_case(text: str) -> str:
    subs = {
        "api": "API",
@@ -340,22 +364,21 @@ def update_nav_file(examples: list[Example]):
            examples_by_category[category] = []
        examples_by_category[category].append(example)

-    # Build the new Examples section - start with preserved items
-    examples_section = preserved_items.copy()
-
-    # Add examples grouped by category, sorted by category name
-    for category in sorted(examples_by_category.keys()):
-        category_examples = sorted(examples_by_category[category], key=lambda e: e.path.stem)
-        category_items = []
-        for example in category_examples:
+    def category_items(category: str) -> list[dict[str, str]]:
+        items = []
+        for example in sorted(examples_by_category[category], key=lambda e: e.path.stem):
            doc_path = EXAMPLE_DOC_DIR / example.category / f"{example.path.stem}.md"
            rel_path = doc_path.relative_to(ROOT_DIR / "docs")
-            category_items.append({example.nav_title: rel_path.as_posix()})
+            items.append({example.nav_title: rel_path.as_posix()})
+        return items

-        if category_items:
-            # Format category name (e.g., "offline_inference" -> "Offline Inference")
+    # Build the new Examples section - start with preserved items.
+    examples_section = preserved_items.copy()
+    for category in sorted(examples_by_category.keys()):
+        items = category_items(category)
+        if items:
            category_title = fix_case(category.replace("_", " ").title())
-            examples_section.append({category_title: category_items})
+            examples_section.append({category_title: items})

    # Update the nav structure
    nav_list[user_guide_idx]["User Guide"][examples_idx]["Examples"] = examples_section
@@ -390,6 +413,8 @@ def on_startup(command: Literal["build", "gh-deploy", "serve"], dirty: bool):
        for path in category.glob("*/*.md"):
            examples.append(Example(path.parent, category.stem))

+    examples = [example for example in examples if is_general_example(example)]
+
    # Generate the example documentation
    for example in sorted(examples, key=lambda e: e.path.stem):
        example_name = f"{example.path.stem}.md"
@@ -17,84 +17,94 @@ th {
 }
 </style>

-| Architecture | Models | Example HF Models | NVIDIA GPU | AMD GPU | Ascend NPU | Intel GPU |
-|--------------|--------|-------------------|------------|---------|-----|-----------|
-| `Qwen3OmniMoeForConditionalGeneration` | Qwen3-Omni | `Qwen/Qwen3-Omni-30B-A3B-Instruct` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `Qwen2_5OmniForConditionalGeneration` | Qwen2.5-Omni | `Qwen/Qwen2.5-Omni-7B`, `Qwen/Qwen2.5-Omni-3B` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `MingFlashOmniForConditionalGeneration` + `MingImagePipeline` | Ming-flash-omni-2.0 (omni-speech + imagegen<sup>1</sup>) | `Jonathan1909/Ming-flash-omni-2.0` | ✅︎ |   |   |   |
-| `BagelForConditionalGeneration` | BAGEL (DiT-only) | `ByteDance-Seed/BAGEL-7B-MoT` | ✅︎ | ✅︎ | | ✅︎ |
-| `InternVLAA1Pipeline` | InternVLA-A1 | `InternRobotics/InternVLA-A1-3B` | ✅︎ | ✅︎ | | |
-| `Gr00tN1d7Pipeline` | GR00T N1.7 | `nvidia/GR00T-N1.7-3B` | ✅︎ | | | |
-| `HunyuanImage3ForCausalMM` | HunyuanImage3.0 (DiT-only) | `tencent/HunyuanImage-3.0`, `tencent/HunyuanImage-3.0-Instruct` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `QwenImagePipeline` | Qwen-Image | `Qwen/Qwen-Image` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `QwenImagePipeline` | Qwen-Image-2512 | `Qwen/Qwen-Image-2512` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `QwenImageEditPipeline` | Qwen-Image-Edit | `Qwen/Qwen-Image-Edit` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `QwenImageEditPlusPipeline` | Qwen-Image-Edit-2509 | `Qwen/Qwen-Image-Edit-2509` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `QwenImageLayeredPipeline` | Qwen-Image-Layered | `Qwen/Qwen-Image-Layered` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `QwenImageEditPlusPipeline` | Qwen-Image-Edit-2511 | `Qwen/Qwen-Image-Edit-2511` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `GlmImagePipeline` | GLM-Image | `zai-org/GLM-Image` | ✅︎ | ✅︎ | | |
-| `ZImagePipeline` | Z-Image | `Tongyi-MAI/Z-Image-Turbo` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `Krea2Pipeline` | Krea 2 (Raw + Turbo) | `krea/Krea-2-Raw`, `krea/Krea-2-Turbo` | ✅︎ | | | |
-| `WanPipeline` | Wan2.1-T2V, Wan2.2-T2V, Wan2.2-TI2V | `Wan-AI/Wan2.1-T2V-1.3B-Diffusers`, `Wan-AI/Wan2.1-T2V-14B-Diffusers`, `Wan-AI/Wan2.2-T2V-A14B-Diffusers`, `Wan-AI/Wan2.2-TI2V-5B-Diffusers` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `WanImageToVideoPipeline` | Wan2.2-I2V | `Wan-AI/Wan2.2-I2V-A14B-Diffusers` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `Cosmos3OmniDiffusersPipeline` | Cosmos3 T2I, T2V, I2V, V2V, T2V with sound, action policy | `nvidia/Cosmos3-Nano`, `nvidia/Cosmos3-Super` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `Wan22S2VPipeline` | Wan2.2-S2V | `Wan-AI/Wan2.2-S2V-14B` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `Wan22VACEPipeline` | Wan2.1-VACE | `Wan-AI/Wan2.1-VACE-1.3B-diffusers`, `Wan-AI/Wan2.1-VACE-14B-diffusers` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `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` | ✅︎ | ✅︎ | | |
-| `LTX2DistilledPipeline` | LTX-2 distilled two-stage T2V and I2V | `rootonchair/LTX-2-19b-distilled` | ✅︎ | ✅︎ | | |
-| `LingBotVideoPipeline` | LingBot-Video dense and MoE T2V | `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` | ✅︎ | | | |
-| `DreamZeroPipeline` | DreamZero-DROID | `GEAR-Dreams/DreamZero-DROID` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `HeliosPipeline`, `HeliosPyramidPipeline` | Helios | `BestWishYsh/Helios-Base`, `BestWishYsh/Helios-Mid`, `BestWishYsh/Helios-Distilled` | ✅︎ | ✅︎ | ✅︎ | |
-| `MagiHumanPipeline` | MagiHuman | `SII-GAIR/daVinci-MagiHuman-Base-1080p` | ✅︎ | ✅︎ | | |
-| `OvisImagePipeline` | Ovis-Image | `OvisAI/Ovis-Image` | ✅︎ | ✅︎ | | ✅︎ |
-| `LongcatImagePipeline` | LongCat-Image | `meituan-longcat/LongCat-Image` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `LongCatImageEditPipeline` | LongCat-Image-Edit | `meituan-longcat/LongCat-Image-Edit` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `BooguImagePipeline` | Boogu-Image | `Boogu/Boogu-Image-0.1-Base`, `Boogu/Boogu-Image-0.1-Edit` | ✅︎ | | | |
-| `StableDiffusionXLPipeline` | Stable-Diffusion-XL | `stabilityai/stable-diffusion-xl-base-1.0` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `StableDiffusion3Pipeline` | Stable-Diffusion-3 | `stabilityai/stable-diffusion-3.5-medium` | ✅︎ | ✅︎ | | ✅︎ |
-| `CosyVoice3Model` | CosyVoice3 | `FunAudioLLM/Fun-CosyVoice3-0.5B-2512` | ✅︎ | ✅︎ | | ✅︎ |
-| `OmniVoiceModel` | OmniVoice | `k2-fsa/OmniVoice` | ✅︎ | | | |
-| `VoxCPM2TalkerForConditionalGeneration` | VoxCPM2 | `openbmb/VoxCPM2` | ✅︎ | | | |
-| `MammothModa2ForConditionalGeneration` | MammothModa2-Preview | `bytedance-research/MammothModa2-Preview` | ✅︎ | ✅︎ | | |
-| `MammothModa2ForConditionalGeneration` | MammothModa2-Dev (AR-only image understanding) | `bytedance-research/MammothModa2-Dev` | ✅︎ | | | |
-| `Flux2KleinPipeline` | FLUX.2-klein | `black-forest-labs/FLUX.2-klein-4B`, `black-forest-labs/FLUX.2-klein-9B` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `FluxKontextPipeline` | FLUX.1-Kontext-dev | `black-forest-labs/FLUX.1-Kontext-dev` | ✅︎ | ✅︎ | | |
-| `FluxPipeline` | FLUX.1-dev | `black-forest-labs/FLUX.1-dev` | ✅︎ | ✅︎ | | ✅︎ |
-| `FluxPipeline` | FLUX.1-schnell | `black-forest-labs/FLUX.1-schnell` | ✅︎ | ✅︎ | | ✅︎ |
-| `OmniGen2Pipeline` | OmniGen2 | `OmniGen2/OmniGen2` | ✅︎ | ✅︎ | | ✅︎ |
-| `StableAudioPipeline` | Stable-Audio-Open | `stabilityai/stable-audio-open-1.0` | ✅︎ | ✅︎ | | ✅︎ |
-| `SoulXSingerPipeline` | SoulX-Singer (SVS) | `Soul-AILab/SoulX-Singer` | ✅︎ | | | |
-| `SoulXSingerSVCPipeline` | SoulX-Singer-SVC | `Soul-AILab/SoulX-Singer` (`model-svc.pt`) | ✅︎ | | | |
-| `AudioXPipeline` | AudioX | `zhangj1an/AudioX` | ✅︎ | ✅︎ | | |
-| `Qwen3TTSForConditionalGeneration` | Qwen3-TTS-12Hz-1.7B-CustomVoice | `Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `Qwen3TTSForConditionalGeneration` | Qwen3-TTS-12Hz-1.7B-VoiceDesign | `Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `Qwen3TTSForConditionalGeneration` | Qwen3-TTS-12Hz-1.7B-Base | `Qwen/Qwen3-TTS-12Hz-1.7B-Base` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-| `MingTTSForConditionalGeneration` | Ming-omni-tts dense 0.5B | `inclusionAI/Ming-omni-tts-0.5B` | ✅︎ | | | |
-| `GLMTTSForConditionalGeneration` | GLM-TTS | `zai-org/GLM-TTS` | ✅︎ | | | |
-| `MossTTSNanoForCausalLM` | MOSS-TTS-Nano | `OpenMOSS-Team/MOSS-TTS-Nano` | ✅︎ | | | |
-| `MossTTSDelayModel` | MOSS-TTS, MOSS-TTSD, MOSS-SoundEffect, MOSS-VoiceGenerator | `OpenMOSS-Team/MOSS-VoiceGenerator` | ✅︎ | | | |
-| `MossTTSRealtime` | MOSS-TTS-Realtime | `OpenMOSS-Team/MOSS-TTS-Realtime` | ✅︎ | | | |
-| `HiggsAudioV2ForConditionalGeneration` | Higgs-Audio v2 | `bosonai/higgs-audio-v2-generation-3B-base` | ✅︎ | | | |
-| `HiggsMultimodalQwen3ForConditionalGeneration` | Higgs-Audio v3 (TTS) | `bosonai/higgs-audio-v3-tts-4b` | ✅︎ | | | |
-| `IndexTTS2TalkerForConditionalGeneration` | IndexTTS-2 | `IndexTeam/IndexTTS-2` | ✅︎ | | | |
-| `NextStep11Pipeline` | NextStep-1.1 | `stepfun-ai/NextStep-1.1` | ✅︎ | ✅︎ | | ✅︎ |
-| `MiMoAudioModel` | MiMo-Audio-7B-Instruct | `XiaomiMiMo/MiMo-Audio-7B-Instruct` | ✅︎ | ✅︎ | | |
-| `MiMoV2ASRForCausalLM` | MiMo-V2.5-ASR | `XiaomiMiMo/MiMo-V2.5-ASR` | ✅︎ | ✅︎ | | |
-| `Flux2Pipeline` | FLUX.2-dev | `black-forest-labs/FLUX.2-dev` | ✅︎ | ✅︎ | | |
-| `FishSpeechSlowARForConditionalGeneration` | Fish Speech S2 Pro | `fishaudio/s2-pro` | ✅︎ | ✅︎ | | ✅︎ |
-| `DreamIDOmniPipeline` | DreamID-Omni | `XuGuo699/DreamID-Omni` | ✅︎ | ✅︎ | | |
-| `SenseNovaU1Pipeline` | SenseNova-U1 (DiT-only) | `SenseNova/SenseNova-U1-8B-MoT` | ✅︎ | | | |
-| `LancePipeline` | Lance | `bytedance-research/Lance` | ✅︎ | | | |
-| `HunyuanVideo15Pipeline` | HunyuanVideo-1.5-T2V | `hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v`, `hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-720p_t2v` | ✅︎ | ✅︎ | | |
-| `HunyuanVideo15ImageToVideoPipeline` | HunyuanVideo-1.5-I2V | `hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_i2v`, `hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-720p_i2v` | ✅︎ | ✅︎ | | |
-| `VoxtralTTSForConditionalGeneration` | Voxtral TTS | `mistralai/Voxtral-4B-TTS-2603` | ✅︎ | ✅︎ | | |
-| `CovoAudioForConditionalGeneration` | Covo-Audio-Chat | `tencent/Covo-Audio-Chat` | ✅︎ | | | |
-|`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` | ✅︎ | ✅︎ | ✅︎ | ✅︎ |
-|`HiDreamImagePipeline` | HiDream-I1-Full | `HiDream-ai/HiDream-I1-Full` | ✅︎ | ✅︎ | | |
+| Architecture | Models | Example HF Models | NVIDIA GPU | AMD GPU | Ascend NPU | Intel GPU | Recipe |
+|--------------|--------|-------------------|------------|---------|-----|-----------|--------|
+| `Qwen3OmniMoeForConditionalGeneration` | Qwen3-Omni | `Qwen/Qwen3-Omni-30B-A3B-Instruct` | ✅︎ | | | | [Repository](https://github.com/vllm-project/vllm-omni/blob/main/recipes/Qwen/Qwen3-Omni.md) |
+| `Qwen2_5OmniForConditionalGeneration` | Qwen2.5-Omni | `Qwen/Qwen2.5-Omni-7B`, `Qwen/Qwen2.5-Omni-3B` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `MingFlashOmniForConditionalGeneration` + `MingImagePipeline` | Ming-flash-omni-2.0 (omni-speech + imagegen<sup>1</sup>) | `Jonathan1909/Ming-flash-omni-2.0` | ✅︎ |   |   |   | — |
+| `BagelForConditionalGeneration` | BAGEL (DiT-only) | `ByteDance-Seed/BAGEL-7B-MoT` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `InternVLAA1Pipeline` | InternVLA-A1 | `InternRobotics/InternVLA-A1-3B` | ✅︎ | ✅︎ | | | — |
+| `Gr00tN1d7Pipeline` | GR00T N1.7 | `nvidia/GR00T-N1.7-3B` | ✅︎ | | | | — |
+| `HunyuanImage3ForCausalMM` | HunyuanImage3.0 (DiT-only) | `tencent/HunyuanImage-3.0`, `tencent/HunyuanImage-3.0-Instruct` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `QwenImagePipeline` | Qwen-Image | `Qwen/Qwen-Image` | ✅︎ | | | | [Published](https://recipes.vllm.ai/Qwen/Qwen-Image) |
+| `QwenImagePipeline` | Qwen-Image-2512 | `Qwen/Qwen-Image-2512` | ✅︎ | | | | [Repository](https://github.com/vllm-project/vllm-omni/blob/main/recipes/Qwen/Qwen-Image-2512.md) |
+| `QwenImageEditPipeline` | Qwen-Image-Edit | `Qwen/Qwen-Image-Edit` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `QwenImageEditPlusPipeline` | Qwen-Image-Edit-2509 | `Qwen/Qwen-Image-Edit-2509` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `QwenImageLayeredPipeline` | Qwen-Image-Layered | `Qwen/Qwen-Image-Layered` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `QwenImageEditPlusPipeline` | Qwen-Image-Edit-2511 | `Qwen/Qwen-Image-Edit-2511` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `GlmImagePipeline` | GLM-Image | `zai-org/GLM-Image` | ✅︎ | ✅︎ | | | — |
+| `ZImagePipeline` | Z-Image | `Tongyi-MAI/Z-Image-Turbo` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `Krea2Pipeline` | Krea 2 (Raw + Turbo) | `krea/Krea-2-Raw`, `krea/Krea-2-Turbo` | ✅︎ | | | | — |
+| `WanPipeline` | Wan2.1-T2V, Wan2.2-T2V, Wan2.2-TI2V | `Wan-AI/Wan2.1-T2V-1.3B-Diffusers`, `Wan-AI/Wan2.1-T2V-14B-Diffusers`, `Wan-AI/Wan2.2-T2V-A14B-Diffusers`, `Wan-AI/Wan2.2-TI2V-5B-Diffusers` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `WanImageToVideoPipeline` | Wan2.2-I2V | `Wan-AI/Wan2.2-I2V-A14B-Diffusers` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `Cosmos3OmniDiffusersPipeline` | Cosmos3 T2I, T2V, I2V, V2V, T2V with sound, action policy | `nvidia/Cosmos3-Nano`, `nvidia/Cosmos3-Super` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `Wan22S2VPipeline` | Wan2.2-S2V | `Wan-AI/Wan2.2-S2V-14B` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `Wan22VACEPipeline` | Wan2.1-VACE | `Wan-AI/Wan2.1-VACE-1.3B-diffusers`, `Wan-AI/Wan2.1-VACE-14B-diffusers` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `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` | ✅︎ | ✅︎ | | | — |
+| `LTX2DistilledPipeline` | LTX-2 distilled two-stage T2V and I2V | `rootonchair/LTX-2-19b-distilled` | ✅︎ | ✅︎ | | | — |
+| `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` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `HeliosPipeline`, `HeliosPyramidPipeline` | Helios | `BestWishYsh/Helios-Base`, `BestWishYsh/Helios-Mid`, `BestWishYsh/Helios-Distilled` | ✅︎ | ✅︎ | ✅︎ | | — |
+| `MagiHumanPipeline` | MagiHuman | `SII-GAIR/daVinci-MagiHuman-Base-1080p` | ✅︎ | ✅︎ | | | — |
+| `OvisImagePipeline` | Ovis-Image | `OvisAI/Ovis-Image` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `LongcatImagePipeline` | LongCat-Image | `meituan-longcat/LongCat-Image` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `LongCatImageEditPipeline` | LongCat-Image-Edit | `meituan-longcat/LongCat-Image-Edit` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `LongCatVideoAvatarPipeline` | LongCat-Video-Avatar-1.5 A2V/AI2V (native single-speaker and multi-speaker AI2V/AVC) | `meituan-longcat/LongCat-Video-Avatar-1.5` | ✅︎ | | | | [Repository](https://github.com/vllm-project/vllm-omni/blob/main/recipes/meituan-longcat/LongCat-Video-Avatar-1.5.md) |
+| `BooguImagePipeline` | Boogu-Image | `Boogu/Boogu-Image-0.1-Base`, `Boogu/Boogu-Image-0.1-Edit` | ✅︎ | | | | — |
+| `StableDiffusionXLPipeline` | Stable-Diffusion-XL | `stabilityai/stable-diffusion-xl-base-1.0` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `StableDiffusion3Pipeline` | Stable-Diffusion-3 | `stabilityai/stable-diffusion-3.5-medium` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `CosyVoice3Model` | CosyVoice3 | `FunAudioLLM/Fun-CosyVoice3-0.5B-2512` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `OmniVoiceModel` | OmniVoice | `k2-fsa/OmniVoice` | ✅︎ | | | | — |
+| `VoxCPM2TalkerForConditionalGeneration` | VoxCPM2 | `openbmb/VoxCPM2` | ✅︎ | | | | — |
+| `MammothModa2ForConditionalGeneration` | MammothModa2-Preview | `bytedance-research/MammothModa2-Preview` | ✅︎ | ✅︎ | | | — |
+| `MammothModa2ForConditionalGeneration` | MammothModa2-Dev (AR-only image understanding) | `bytedance-research/MammothModa2-Dev` | ✅︎ | | | | — |
+| `Flux2KleinPipeline` | FLUX.2-klein | `black-forest-labs/FLUX.2-klein-4B`, `black-forest-labs/FLUX.2-klein-9B` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `FluxKontextPipeline` | FLUX.1-Kontext-dev | `black-forest-labs/FLUX.1-Kontext-dev` | ✅︎ | ✅︎ | | | — |
+| `FluxPipeline` | FLUX.1-dev | `black-forest-labs/FLUX.1-dev` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `FluxPipeline` | FLUX.1-schnell | `black-forest-labs/FLUX.1-schnell` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `OmniGen2Pipeline` | OmniGen2 | `OmniGen2/OmniGen2` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `StableAudioPipeline` | Stable-Audio-Open | `stabilityai/stable-audio-open-1.0` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `SoulXSingerPipeline` | SoulX-Singer (SVS) | `Soul-AILab/SoulX-Singer` | ✅︎ | | | | — |
+| `SoulXSingerSVCPipeline` | SoulX-Singer-SVC | `Soul-AILab/SoulX-Singer` (`model-svc.pt`) | ✅︎ | | | | — |
+| `AudioXPipeline` | AudioX | `zhangj1an/AudioX` | ✅︎ | ✅︎ | | | — |
+| `Qwen3TTSForConditionalGeneration` | Qwen3-TTS-12Hz-1.7B-CustomVoice | `Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice` | ✅︎ | | | | [Published](https://recipes.vllm.ai/Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice) |
+| `Qwen3TTSForConditionalGeneration` | Qwen3-TTS-12Hz-1.7B-VoiceDesign | `Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `Qwen3TTSForConditionalGeneration` | Qwen3-TTS-12Hz-1.7B-Base | `Qwen/Qwen3-TTS-12Hz-1.7B-Base` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+| `MingTTSForConditionalGeneration` | Ming-omni-tts dense 0.5B | `inclusionAI/Ming-omni-tts-0.5B` | ✅︎ | | | | — |
+| `GLMTTSForConditionalGeneration` | GLM-TTS | `zai-org/GLM-TTS` | ✅︎ | | | | — |
+| `MossTTSNanoForCausalLM` | MOSS-TTS-Nano | `OpenMOSS-Team/MOSS-TTS-Nano` | ✅︎ | | | | — |
+| `MossTTSDelayModel` | MOSS-TTS, MOSS-TTSD, MOSS-SoundEffect, MOSS-VoiceGenerator | `OpenMOSS-Team/MOSS-VoiceGenerator` | ✅︎ | | | | — |
+| `MossTTSRealtime` | MOSS-TTS-Realtime | `OpenMOSS-Team/MOSS-TTS-Realtime` | ✅︎ | | | | — |
+| `HiggsAudioV2ForConditionalGeneration` | Higgs-Audio v2 | `bosonai/higgs-audio-v2-generation-3B-base` | ✅︎ | | | | — |
+| `HiggsMultimodalQwen3ForConditionalGeneration` | Higgs-Audio v3 (TTS) | `bosonai/higgs-audio-v3-tts-4b` | ✅︎ | | | | [Published](https://recipes.vllm.ai/bosonai/higgs-audio-v3-tts-4b) |
+| `IndexTTS2TalkerForConditionalGeneration` | IndexTTS-2 | `IndexTeam/IndexTTS-2` | ✅︎ | | | | — |
+| `NextStep11Pipeline` | NextStep-1.1 | `stepfun-ai/NextStep-1.1` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `MiMoAudioModel` | MiMo-Audio-7B-Instruct | `XiaomiMiMo/MiMo-Audio-7B-Instruct` | ✅︎ | ✅︎ | | | — |
+| `MiMoV2ASRForCausalLM` | MiMo-V2.5-ASR | `XiaomiMiMo/MiMo-V2.5-ASR` | ✅︎ | ✅︎ | | | — |
+| `Flux2Pipeline` | FLUX.2-dev | `black-forest-labs/FLUX.2-dev` | ✅︎ | ✅︎ | | | — |
+| `FishSpeechSlowARForConditionalGeneration` | Fish Speech S2 Pro | `fishaudio/s2-pro` | ✅︎ | ✅︎ | | ✅︎ | — |
+| `DreamIDOmniPipeline` | DreamID-Omni | `XuGuo699/DreamID-Omni` | ✅︎ | ✅︎ | | | — |
+| `SenseNovaU1Pipeline` | SenseNova-U1 (DiT-only) | `SenseNova/SenseNova-U1-8B-MoT` | ✅︎ | | | | — |
+| `LancePipeline` | Lance | `bytedance-research/Lance` | ✅︎ | | | | — |
+| `HunyuanVideo15Pipeline` | HunyuanVideo-1.5-T2V | `hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v`, `hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-720p_t2v` | ✅︎ | ✅︎ | | | — |
+| `HunyuanVideo15ImageToVideoPipeline` | HunyuanVideo-1.5-I2V | `hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_i2v`, `hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-720p_i2v` | ✅︎ | ✅︎ | | | — |
+| `VoxtralTTSForConditionalGeneration` | Voxtral TTS | `mistralai/Voxtral-4B-TTS-2603` | ✅︎ | ✅︎ | | | — |
+| `CovoAudioForConditionalGeneration` | Covo-Audio-Chat | `tencent/Covo-Audio-Chat` | ✅︎ | | | | — |
+|`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` | ✅︎ | ✅︎ | ✅︎ | ✅︎ | — |
+|`HiDreamImagePipeline` | HiDream-I1-Full | `HiDream-ai/HiDream-I1-Full` | ✅︎ | ✅︎ | | | — |
+
+The `Recipe` column links the corresponding deployment recipe. `Published` is
+preferred when a corresponding `recipes.vllm.ai` page exists; otherwise the
+link points to the repository recipe. `—` means that no recipe link has been
+added for that model row yet.
+
+For rows with an audited recipe link, the hardware checkmarks reflect hardware
+explicitly documented in that recipe. Other rows retain the implementation
+support metadata until their recipe is audited.

 ✅︎ indicates the model is supported on that backend. Empty cells mean not listed as supported on that backend.

@@ -283,23 +283,20 @@ blocks — use Ulysses SP (`ring_degree=1`).

 ### Which geometries run sparse

-Sparsity is only applied when the video segment is a **multiple of 128 rows**, where the row count
-is `latent_t × (height / 32) × (width / 32)`. Otherwise `rf_v2`'s block mask and the kernel's own
-tiling disagree on the block count, and the block straddling the seam mixes video and prefix rows
-that selection may then drop. Any resolution still runs — an unaligned geometry falls back to dense
-attention and logs `RAINFUSION_ATTN staying dense` with the row count it computed — but it gets no
-speedup, so pick an aligned geometry when you want one.
+MindIE-SD rf_v2 handles arbitrary video grids by spatially rearranging video first and, when
+necessary, promoting a real-video suffix to its always-kept prefix segment. It never pads: rf_v2
+does not consume an attention mask for padded keys. The remaining video rows stay sparse, while the
+promoted rows remain visible to every query and key.

-Alignment is necessary, not sufficient, for good quality. `rf_v2` groups video positions into 8x8
-spatial tiles (two tiles fill one 128-row block), which is what makes a selected block a compact
-patch of the frame. When the latent `h` or `w` is not a multiple of 8, the leftover rows or columns
-are peeled off and appended as a flat run instead of being tiled, so they get pooled with spatially
-distant positions and selection can no longer rank them meaningfully. This is silent, and invisible
-to a `sparsity=0` check, because a fully populated mask does not care how blocks are grouped.
+For example, MiniMax-H3 at 1344x768 has grid `(62, 24, 42)`: its 62496 video rows leave a 32-row
+128-block residue, and its two non-8-wide latent columns form a 2928-row rearranged suffix.
+MindIE-SD promotes 2976 rows, leaving 59520 sparse rows (`465 × 128`). This keeps the block mask
+and kernel tiling aligned, at the cost of retaining more blocks.

-"Multiple of 8" on the latent grid means **width and height that are multiples of 256**. Aligned
-resolutions off that grid still run sparse; they just lose more fidelity at the same `sparsity`,
-which you can buy back with `start_step`.
+8x8-aligned spatial grids remain preferable for performance. A latent `h` or `w` not divisible by 8
+creates an always-kept residual suffix and lowers realized sparsity. "Multiple of 8" on the latent
+grid means **width and height that are multiples of 256**. This path requires a MindIE-SD release
+that implements protected video tails for rf_v2.

 ## End-to-End Benchmark (BF16, sm_120 RTX Pro 6000 Blackwell)

@@ -1,14 +1,18 @@
 # LoRA (Low-Rank Adaptation) Guide

-LoRA (Low-Rank Adaptation) enables fine-tuning diffusion models by adding trainable low-rank matrices to existing model weights. vLLM-Omni currently supports PEFT-style LoRA adapters, allowing you to customize model behavior without modifying the base model weights.
+LoRA (Low-Rank Adaptation) enables fine-tuning diffusion models by adding trainable low-rank matrices to existing model weights. vLLM-Omni supports two LoRA backends: **PEFT** for PEFT-style adapters, and **Distill** usually for few-steps inference. **PEFT** backend allows you to customize model behavior without modifying the base model weights. **Distill** backend fuses LoRA weights into the base model at initialization.

 ## Overview

-LoRA adapters are lightweight, model-specific fine-tuning weights that can be dynamically loaded and applied to diffusion models. vLLM-Omni uses a unified LoRA handling mechanism similar to vLLM with LRU cache management.
+LoRA adapters are lightweight, model-specific fine-tuning weights that can be applied to diffusion models in two ways:
+
+- **PEFT backend** (`--lora-backend peft`, default): Loads a PEFT-format adapter folder via `DiffusionLoRAManager`. Adapters are cached (LRU) and activated per request via `LoRARequest`. It uses a unified LoRA handling mechanisms similar to vLLM with LRU cache management.
+- **Distill backend** (`--lora-backend distill`): Calls `pipeline.load_lora_weights` once at initialization to fuse one or more concrete checkpoint files directly into the base weights. Typically used for distilled few-step LoRAs (e.g. Lightning, LightX2V).

 ## LoRA Adapter Format

-LoRA adapters must be in **PEFT (Parameter-Efficient Fine-Tuning)** format. A typical LoRA adapter directory structure:
+### PEFT (Parameter-Efficient Fine-Tuning) format (default)
+A typical PEFT-format LoRA adapter directory structure:

 ```
 lora_adapter/
@@ -25,19 +29,20 @@ The `adapter_config.json` file contains metadata about the LoRA adapter, includi

 ### Offline Inference

-#### Pre-loaded LoRA
+#### PEFT backend: pre-loaded LoRA

-Load a LoRA adapter at initialization. This adapter is pre-loaded into the cache and can be activated by requests:
+Load a PEFT-format LoRA adapter at initialization. The adapter is pre-loaded into the cache and can be activated per request:

 ```python
 from vllm_omni import Omni
 from vllm_omni.lora.request import LoRARequest

-lora_path="/path/to/lora_adapter"
+lora_path = "/path/to/lora_adapter"

 omni = Omni(
    model="stabilityai/stable-diffusion-3.5-medium",
-    lora_path=lora_path
+    lora_path=lora_path,
+    lora_backend="peft",  # default, can be omitted
 )

 lora_request = LoRARequest(
@@ -53,6 +58,98 @@ outputs = omni.generate(
 )
 ```

+#### Distill backend: fuse distilled LoRA at init
+
+For distilled few-step LoRAs, pass `lora_backend="distill"` together with one or more concrete `.safetensors` files. The weights are fused into the base model once at init; subsequent `generate()` calls do not need a `LoRARequest`.
+
+##### Supported pipelines
+
+For Qwen-Image and Wan pipelines, the distill backend calls `pipeline.load_lora_weights(...)` during worker initialization.
+
+| Pipeline | Supported distilled LoRA repo | Notes |
+|----------|--------------------------------|-------|
+| `QwenImagePipeline` | `lightx2v/Qwen-Image-2512-Lightning` | Used with Qwen-Image-2512 Lightning-style few-step inference. |
+| `Wan22Pipeline` | `lightx2v/Wan2.1-Distill-Loras`, `lightx2v/Wan2.2-Distill-Loras` | Wan2.1 uses one LoRA file. For dual-transformer Wan2.2 MoE, pass high-noise then low-noise LoRA files. |
+| `Wan22I2VPipeline` | `lightx2v/Wan2.2-Distill-Loras` | For dual-transformer Wan2.2 MoE, pass high-noise then low-noise LoRA files. |
+
+Other diffusion pipelines are not currently listed as supporting distilled LoRA. Use the PEFT backend for request-time adapters, or bake converted weights into a local Diffusers directory before serving.
+
+Single-file example (Qwen-Image-Lightning):
+
+```python
+from vllm_omni import Omni
+
+omni = Omni(
+    model="Qwen/Qwen-Image-2512",
+    lora_path="/path/to/Qwen-Image-2512-Lightning.safetensors",
+    lora_backend="distill",
+)
+
+outputs = omni.generate(prompt="A piece of cheesecake")
+```
+
+Multi-file example (Wan2.2 MoE, high + low noise):
+
+```python
+from vllm_omni import Omni
+
+omni = Omni(
+    model="Wan-AI/Wan2.2-T2V-A14B-Diffusers",
+    lora_path=[
+        "/path/to/wan2.2_high_noise_lora.safetensors",   # -> transformer
+        "/path/to/wan2.2_low_noise_lora.safetensors",    # -> transformer_2
+    ],
+    lora_backend="distill",
+)
+```
+
+The CLI examples under `examples/offline_inference/` accept the same flags, e.g.:
+
+```bash
+python examples/offline_inference/text_to_video/text_to_video.py \
+  --model Wan-AI/Wan2.2-T2V-A14B-Diffusers \
+  --lora-backend distill \
+  --lora-path /path/to/high.safetensors /path/to/low.safetensors \
+  --prompt "A cat playing with yarn"
+```
+
+##### Online serving
+
+Distilled LoRAs can also be fused when an online diffusion server starts. The
+adapter remains active for the lifetime of that server, so requests do not pass
+a per-request `lora` field.
+
+```bash
+vllm serve Qwen/Qwen-Image-2512 \
+  --omni \
+  --port 8091 \
+  --lora-backend distill \
+  --lora-path /path/to/Qwen-Image-2512-Lightning-4steps.safetensors
+```
+
+Then send a normal generation request with the sampling settings expected by
+the distilled checkpoint:
+
+```bash
+curl -s http://localhost:8091/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "messages": [
+      {"role": "user", "content": "A Chinese female college student, around 20 years old, with a very short haircut that conveys a gentle, artistic vibe. Her hair naturally falls to partially cover her cheeks, projecting a tomboyish yet charming demeanor. She has cool-toned fair skin and delicate features, with a slightly shy yet subtly confident expression—her mouth crooked in a playful, youthful smirk. She wears an off-shoulder top, revealing one shoulder, with a well-proportioned figure. The image is framed as a close-up selfie: she dominates the foreground, while the background clearly shows her dormitory—a neatly made bed with white linens on the top bunk, a tidy study desk with organized stationery, and wooden cabinets and drawers. The photo is captured on a smartphone under soft, even ambient lighting, with natural tones, high clarity, and a bright, lively atmosphere full of youthful, everyday energy."}
+    ],
+    "extra_body": {
+      "height": 1024,
+      "width": 1024,
+      "num_inference_steps": 4,
+      "true_cfg_scale": 1.0,
+      "seed": 42
+    }
+  }' | jq -r '.choices[0].message.content[0].image_url.url' | cut -d',' -f2 | base64 -d > qwen_image_distill_lora_output.png
+```
+
+For Wan2.2 MoE serving, pass the high-noise checkpoint first and the low-noise
+checkpoint second to `--lora-path`.
+
 !!! note "Server-side Path Requirement"
    The LoRA adapter path (`local_path`) must be readable on the **server** machine. If your client and server are on different machines, ensure the LoRA adapter is accessible via a shared mount or copied to the server.

@@ -141,6 +238,7 @@ Notes:

 - This route avoids runtime LoRA loading changes in vLLM-Omni when you choose to bake converted weights into a local Diffusers directory.
 - Output quality and speed depend on the replacement checkpoints and sampling params you choose.
+- If you only need to fuse distilled LoRAs into a Wan2.2 checkpoint at load time (without the full LightX2V convert + assemble pipeline), you can instead pass them directly via `--lora-backend distill --lora-path <high>.safetensors <low>.safetensors`. See the [Distill backend](#distill-backend-fuse-distilled-lora-at-init) section above.


 ## See Also
@@ -164,10 +164,12 @@ In `DiffusionParallelConfig`:
 | `ulysses_degree` | int | 1 | Number of GPUs for Ulysses-SP. Uses all-to-all communication. |
 | `ring_degree` | int | 1 | Number of GPUs for Ring-Attention. Uses P2P ring communication. |
 | `ulysses_mode` | str | `"default"` | Ulysses attention mode. Set to `"advanced_uaa"` to handle arbitrary sequence lengths and head counts without padding. |
+| `mask_sp_padding` | bool | `False` | When the sequence length is not divisible by the SP size, tokens are auto-padded with zeros. Set to `True` to mask those padding tokens (strict, but uses the slower varlen attention path); the default `False` leaves them unmasked, keeping the fast path with negligible numerical impact. |

 **Notes:**
 - Total sequence parallel size equals to `ulysses_degree × ring_degree`
 - Degrees must evenly divide the sequence length for optimal performance (or use `ulysses_mode="advanced_uaa"` for Ulysses-SP)
+- `mask_sp_padding` is an experimental feature, currently only supported by `Wan2.2`, `Wan2.2 Vace`, `Qwen-Image`, `Flux 2`, and `HunyuanVideo 1.5`


 ## Best Practices
@@ -159,7 +159,8 @@ The following tables show which models support each feature:
 | **HunyuanVideo-1.5 T2V I2V** |     ❌     |     ✅      |           ✅           |       ✅        |         ✅         |         ❌         |   ✅    |             ✅             |  ✅ (encode/decode)   |       ✅        |        ❌         |
 | **DreamID-Omni**             |     ❌     |     ❌      |           ❌           |       ✅        |         ❌         |         ❌         |   ✅    |             ✅             |          ❌           |       ❌        |        ❌         |
 | **Cosmos3**                  |     ❌     |     ✅      |           ✅           |       ✅        |         ✅         |         ❌         |   ✅    |             ✅             |  ✅ (encode/decode)   |       ✅        |        ❌         |
-| **MiniMax-H3**               |     ❌     |     ✅      |           ✅           |       ❌        |       ✅ (DiT/TE)  |         ❌         |   ✅    |             ✅             |       ✅ (tile)       |      ✅ (DiT)      |        ❌         |
+| **LongCat-Video-Avatar-1.5** |     ❌     |     ❌      |           ❌           |       ❌        |         ❌         |         ❌         |   ❌    |             ❌             |          ❌           |       ❌        |        ❌         |
+| **MiniMax-H3**               | ✅ (FL2VA) |     ✅      |           ✅           |       ❌        |       ✅ (DiT/TE)  |         ❌         |   ✅    |             ✅             |       ✅ (tile)       |      ✅ (DiT)      |        ❌         |

 **Frame Interpolation Support**

@@ -88,8 +88,9 @@ python text_to_image.py \
 | `--vae-use-tiling` | flag | off | Enable VAE tiling for memory optimization |
 | `--cfg-parallel-size` | int | `1` | Set to `2` to enable CFG Parallel |
 | `--enable-cpu-offload` | flag | off | Enable CPU offloading for diffusion models |
-| `--lora-path` | str | — | Path to PEFT LoRA adapter folder |
+| `--lora-path` | str | — | Path to PEFT LoRA adapter folder or checkpoint file |
 | `--lora-scale` | float | `1.0` | Scale factor for LoRA weights |
+| `--lora-backend` | str |`"peft"`| LoRA backend for loading LoRA adapters. Default: peft. Available options: peft, distill |

 **NextStep-1.1 specific arguments:**

@@ -125,6 +125,9 @@ python text_to_video.py \
 - `--audio-sample-rate`: fallback audio sample rate when the pipeline returns audio.
 - `--quantization`: quantization method (`fp8` for FP8, `gguf` for GGUF).
 - `--flow-shift`: scheduler flow_shift parameter.
+- `--lora-path`: path to PEFT LoRA adapter folder or checkpoint file.
+- `--lora-scale`: scale factor for LoRA weights.
+- `--lora-backend`: backend for loading LoRA adapters. Default: peft. Available options: peft, distill.

 ### Wan2.2-specific

@@ -99,7 +99,7 @@ a native vLLM-Omni pipeline for those cases.
 ### Attention Backends

 The diffusers backend converts
-[vLLM-Omni standard of attention backend setting](../../../docs/user_guide/diffusion/attention_backends.md)
+[vLLM-Omni standard of attention backend setting](../../diffusion/attention_backends.md)
 to [diffusers standard](https://huggingface.co/docs/diffusers/optimization/attention_backends#available-backends).

 Specifically for `FLASH_ATTN`, it will first attempt to use FlashAttention-3 and then FlashAttention-2.
@@ -33,6 +33,25 @@ This folder provides a unified CLI script for image-to-video generation using vL

 Default model: `Wan-AI/Wan2.2-I2V-A14B-Diffusers`.

+### LingBot-Video TI2V
+
+LingBot-Video accepts exactly one first-frame image and uses the same dense or
+MoE checkpoint as its T2I and T2V modes:
+
+```bash
+python examples/offline_inference/image_to_video/image_to_video.py \
+  --model robbyant/lingbot-video-dense-1.3b \
+  --model-class-name LingBotVideoPipeline \
+  --image /path/to/input.png \
+  --prompt "the fox looks toward the camera" \
+  --height 192 --width 320 --num-frames 9 --num-inference-steps 2 \
+  --guidance-scale 3.0 --flow-shift 3.0 --fps 24 \
+  --output lingbot_ti2v.mp4
+```
+
+The input image keeps its original geometry until the LingBot pipeline applies
+its model-specific resize and center crop.
+
 ## Prerequisites

 Download the example image used in the snippets below:
@@ -60,6 +60,8 @@ 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.lora.request import LoRARequest
+from vllm_omni.lora.utils import stable_lora_int_id
 from vllm_omni.model_extras import (
    build_image_to_video_prompt,
    get_extra_body_params,
@@ -294,6 +296,30 @@ def parse_args() -> argparse.Namespace:
        default=None,
        help='JSON profiler config for torch/cuda profiling, e.g. \'{"profiler":"torch","torch_profiler_dir":"./perf"}\'.',
    )
+    parser.add_argument(
+        "--lora-path",
+        type=str,
+        nargs="+",
+        default=None,
+        help="Path to a LoRA adapter folder (PEFT format) or concrete LoRA checkpoint files. "
+        "For Wan2.2 MoE models, pass the high-noise checkpoint first and the low-noise checkpoint second.",
+    )
+    parser.add_argument(
+        "--lora-scale",
+        type=float,
+        default=1.0,
+        help="Scale factor for PEFT LoRA weights (default: 1.0).",
+    )
+    parser.add_argument(
+        "--lora-backend",
+        type=str,
+        default="peft",
+        choices=["peft", "distill"],
+        help=(
+            "LoRA backend. 'peft' loads a PEFT-format adapter for request-time activation; "
+            "'distill' fuses one or more concrete LoRA checkpoint files at initialization."
+        ),
+    )
    parser.add_argument(
        "--extra-body",
        type=functools.partial(parse_json_object, flag_name="--extra-body"),
@@ -487,6 +513,12 @@ def main():
        omni_kwargs["flow_shift"] = flow_shift
    if args.quantization is not None:
        omni_kwargs["quantization"] = args.quantization
+    if args.lora_path is not None:
+        lora_path = args.lora_path
+        if len(lora_path) == 1:
+            lora_path = lora_path[0]
+        omni_kwargs["lora_path"] = lora_path
+        omni_kwargs["lora_backend"] = args.lora_backend
    # Cosmos3 loads its (gated) guardrail models at build time, so the guardrails
    # gate is an engine-level config (offline analog of the server's --no-guardrails).
    if args.extra_body and "guardrails" in args.extra_body:
@@ -517,6 +549,18 @@ def main():
    print(f"  Video size: {width}x{height}")
    print(f"{'=' * 60}\n")

+    lora_request = None
+    if args.lora_path and args.lora_backend == "peft":
+        if len(args.lora_path) != 1:
+            raise ValueError("Only one LoRA path is expected for PEFT backend.")
+
+        lora_path = args.lora_path[0]
+        lora_request = LoRARequest(
+            lora_name=Path(lora_path).stem,
+            lora_int_id=stable_lora_int_id(lora_path),
+            lora_path=lora_path,
+        )
+
    negative_prompt = args.negative_prompt
    if negative_prompt is None and not is_ltx2:
        # Preserve the historical empty-prompt behavior for non-LTX examples.
@@ -530,6 +574,11 @@ def main():
        width=width,
        num_frames=num_frames,
    )
+    sampling_extra_args = {"sample_solver": args.sample_solver}
+    if lora_request is not None:
+        sampling_extra_args["lora_request"] = lora_request
+        sampling_extra_args["lora_scale"] = args.lora_scale
+
    sampling_params = OmniDiffusionSamplingParams(
        height=height,
        width=width,
@@ -540,7 +589,7 @@ def main():
        num_inference_steps=num_inference_steps,
        num_frames=num_frames,
        frame_rate=frame_rate,
-        extra_args={"sample_solver": args.sample_solver},
+        extra_args=sampling_extra_args,
    )
    if flow_shift is not None:
        sampling_params.extra_args["flow_shift"] = flow_shift
@@ -0,0 +1,319 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+import argparse
+import json
+import re
+import subprocess
+import tempfile
+import wave
+from functools import lru_cache
+from pathlib import Path
+from typing import Any
+
+import imageio
+import numpy as np
+import torch
+
+from vllm_omni.diffusion.data import DiffusionParallelConfig
+from vllm_omni.diffusion.models.longcat_video.pipeline_longcat_video_avatar import (
+    prepare_longcat_video_avatar_model_for_omni,
+)
+from vllm_omni.entrypoints.omni import Omni
+from vllm_omni.inputs.data import OmniDiffusionSamplingParams
+from vllm_omni.outputs import OmniRequestOutput
+from vllm_omni.platforms import current_omni_platform
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="Generate LongCat-Video-Avatar A2V/AI2V video.")
+    parser.add_argument("--model", default="meituan-longcat/LongCat-Video-Avatar-1.5")
+    parser.add_argument("--base-model-dir", default=None, help="Optional local LongCat-Video base model directory.")
+    parser.add_argument("--stage", choices=["at2v", "ai2v"], default="at2v")
+    parser.add_argument("--prompt", default="A person speaks calmly while facing the camera.")
+    parser.add_argument("--negative-prompt", default="low quality, blurry, watermark, text")
+    parser.add_argument("--audio", default=None, help="Input speech audio path.")
+    parser.add_argument("--image", default=None, help="Input reference image path for AI2V.")
+    parser.add_argument("--input-json", default=None, help="Official LongCat Avatar JSON case for AI2V.")
+    parser.add_argument("--output", default="longcat_avatar_output.mp4")
+    parser.add_argument("--resolution", choices=["480p", "720p"], default="480p")
+    parser.add_argument("--num-frames", type=int, default=93)
+    parser.add_argument("--num-segments", default="1", help="Number of AVC segments, or 'auto' to cover full audio.")
+    parser.add_argument("--num-cond-frames", type=int, default=13)
+    parser.add_argument("--ref-img-index", type=int, default=10)
+    parser.add_argument("--mask-frame-range", type=int, default=3)
+    parser.add_argument("--fps", type=int, default=25)
+    parser.add_argument("--num-inference-steps", type=int, default=8)
+    parser.add_argument("--seed", type=int, default=42)
+    parser.add_argument("--use-distill", action=argparse.BooleanOptionalAction, default=True)
+    parser.add_argument("--use-int8", action=argparse.BooleanOptionalAction, default=True)
+    parser.add_argument(
+        "--build-components-on-gpu",
+        action="store_true",
+        default=False,
+        help="Build large Avatar components directly on GPU for faster startup. Requires more peak VRAM.",
+    )
+    parser.add_argument("--use-kv-cache", action=argparse.BooleanOptionalAction, default=True)
+    parser.add_argument("--offload-kv-cache", action=argparse.BooleanOptionalAction, default=False)
+    parser.add_argument("--enforce-eager", action="store_true")
+    return parser.parse_args()
+
+
+def _extract_frames(output: Any):
+    if isinstance(output, list):
+        output = output[0] if output else None
+    if isinstance(output, OmniRequestOutput):
+        if not output.images:
+            raise ValueError("No video frames found in OmniRequestOutput.")
+        frames = output.images[0] if len(output.images) == 1 else output.images
+        if isinstance(frames, tuple) and len(frames) == 2:
+            frames = frames[0]
+        if isinstance(frames, dict):
+            frames = frames.get("frames") or frames.get("video")
+    else:
+        frames = output
+    if frames is None:
+        raise ValueError("No video frames found in output.")
+    return frames
+
+
+def _frames_to_uint8(frames):
+    frames_np = np.asarray([np.asarray(frame) for frame in frames])
+    if frames_np.ndim == 5:
+        frames_np = frames_np[0]
+    if np.issubdtype(frames_np.dtype, np.floating):
+        frames_np = (np.clip(frames_np, 0.0, 1.0) * 255).round()
+    return np.clip(frames_np, 0, 255).astype(np.uint8)
+
+
+@lru_cache(maxsize=1)
+def _ffmpeg_exe() -> str:
+    try:
+        import imageio_ffmpeg
+
+        return imageio_ffmpeg.get_ffmpeg_exe()
+    except Exception as exc:
+        raise RuntimeError("LongCat Avatar example requires the Python imageio[ffmpeg] dependency.") from exc
+
+
+def _wav_duration(path: Path) -> float:
+    with wave.open(str(path), "rb") as audio:
+        return audio.getnframes() / float(audio.getframerate())
+
+
+def _save_video_with_audio(frames, output_path: Path, audio_path: str, fps: int) -> None:
+    frames_np = _frames_to_uint8(frames)
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+    with tempfile.TemporaryDirectory(prefix="longcat_avatar_export_") as tmp:
+        tmp_dir = Path(tmp)
+        temp_video = tmp_dir / "video.mp4"
+        crop_audio = tmp_dir / "audio.wav"
+        crop_video = tmp_dir / "video_cropped.mp4"
+        writer = imageio.get_writer(str(temp_video), fps=fps, quality=5)
+        try:
+            for frame in frames_np:
+                writer.append_data(frame)
+        finally:
+            writer.close()
+
+        duration = len(frames_np) / fps
+        ffmpeg = _ffmpeg_exe()
+        subprocess.run([ffmpeg, "-y", "-i", audio_path, "-t", f"{duration}", str(crop_audio)], check=True)
+        audio_duration = _wav_duration(crop_audio)
+        subprocess.run(
+            [
+                ffmpeg,
+                "-y",
+                "-i",
+                str(temp_video),
+                "-t",
+                f"{audio_duration}",
+                "-c:v",
+                "copy",
+                str(crop_video),
+            ],
+            check=True,
+        )
+        subprocess.run(
+            [
+                ffmpeg,
+                "-y",
+                "-i",
+                str(crop_video),
+                "-i",
+                str(crop_audio),
+                "-c:v",
+                "libx264",
+                "-c:a",
+                "aac",
+                "-shortest",
+                str(output_path),
+            ],
+            check=True,
+        )
+
+
+def _infer_asset_root_from_input_json(input_json: Path) -> Path:
+    resolved = input_json.resolve()
+    parts = resolved.parts
+    if "assets" not in parts:
+        return resolved.parent
+    asset_idx = parts.index("assets")
+    return Path(*parts[:asset_idx])
+
+
+def _resolve_case_path(path: str, asset_root: Path) -> str:
+    raw_path = Path(path)
+    if raw_path.is_absolute():
+        return str(raw_path)
+    candidate = asset_root / raw_path
+    return str(candidate if candidate.exists() else raw_path)
+
+
+def _speaker_sort_key(name: str) -> tuple[int, str]:
+    match = re.search(r"(\d+)$", str(name))
+    return (int(match.group(1)) if match else 10_000, str(name))
+
+
+def _load_input_case(input_json: str) -> tuple[dict[str, Any], Path]:
+    input_json_path = Path(input_json)
+    with input_json_path.open("r", encoding="utf-8") as f:
+        case = json.load(f)
+    return case, _infer_asset_root_from_input_json(input_json_path)
+
+
+def _create_merged_audio_for_case(
+    audio_paths: dict[str, str],
+    audio_type: str,
+    output_path: Path,
+) -> str:
+    sorted_paths = [path for _, path in sorted(audio_paths.items(), key=lambda item: _speaker_sort_key(str(item[0])))]
+    if len(sorted_paths) == 1:
+        return sorted_paths[0]
+    inputs = [arg for path in sorted_paths for arg in ("-i", path)]
+    audio_type = audio_type.lower()
+    if audio_type == "para":
+        filter_complex = f"amix=inputs={len(sorted_paths)}:duration=longest:normalize=0"
+    elif audio_type == "add":
+        filter_complex = "".join(f"[{idx}:a]" for idx in range(len(sorted_paths)))
+        filter_complex += f"concat=n={len(sorted_paths)}:v=0:a=1"
+    else:
+        raise NotImplementedError(f"Unsupported LongCat Avatar audio_type {audio_type!r}.")
+    subprocess.run([_ffmpeg_exe(), "-y", *inputs, "-filter_complex", filter_complex, str(output_path)], check=True)
+    return str(output_path)
+
+
+def main():
+    args = parse_args()
+    input_case = None
+    input_asset_root = None
+    audio_for_mux = args.audio
+    if args.input_json:
+        input_case, input_asset_root = _load_input_case(args.input_json)
+        args.stage = "ai2v"
+        args.prompt = input_case.get("prompt") or args.prompt
+        args.image = _resolve_case_path(input_case["cond_image"], input_asset_root)
+        args.audio = None
+    if not args.input_json and not args.audio:
+        raise ValueError("--audio is required unless --input-json is provided.")
+    if args.stage == "ai2v" and not args.image:
+        raise ValueError("--image is required when --stage=ai2v.")
+    model_for_omni = prepare_longcat_video_avatar_model_for_omni(args.model, args.use_int8)
+
+    additional_config = {
+        "model_type": "avatar-v1.5",
+        "resolution": args.resolution,
+        "use_distill": args.use_distill,
+        "use_int8": args.use_int8,
+        "build_components_on_gpu": args.build_components_on_gpu,
+    }
+    if args.base_model_dir:
+        additional_config["base_model_dir"] = args.base_model_dir
+
+    omni = Omni(
+        model=model_for_omni,
+        model_class_name="LongCatVideoAvatarPipeline",
+        dtype="bfloat16",
+        enforce_eager=args.enforce_eager,
+        additional_config=additional_config,
+        parallel_config=DiffusionParallelConfig(
+            ulysses_degree=1,
+            ring_degree=1,
+            cfg_parallel_size=1,
+            tensor_parallel_size=1,
+            vae_patch_parallel_size=1,
+            pipeline_parallel_size=1,
+        ),
+    )
+
+    audio_input: str | dict[str, str] | None = args.audio
+    additional_information = {"stage": args.stage, "resolution": args.resolution}
+    if input_case is not None and input_asset_root is not None:
+        audio_input = {
+            speaker: _resolve_case_path(path, input_asset_root)
+            for speaker, path in sorted(
+                input_case["cond_audio"].items(),
+                key=lambda item: _speaker_sort_key(str(item[0])),
+            )
+        }
+        additional_information.update(
+            {
+                "input_json": args.input_json,
+                "audio_type": input_case.get("audio_type", "para"),
+                "bbox": input_case.get("bbox", {}),
+            }
+        )
+
+    prompt = {
+        "prompt": args.prompt,
+        "negative_prompt": args.negative_prompt,
+        "multi_modal_data": {"audio": audio_input},
+        "additional_information": additional_information,
+    }
+    if args.stage == "ai2v":
+        prompt["multi_modal_data"]["image"] = args.image
+
+    try:
+        output = omni.generate(
+            prompt,
+            OmniDiffusionSamplingParams(
+                generator=torch.Generator(device=current_omni_platform.device_type).manual_seed(args.seed),
+                seed=args.seed,
+                guidance_scale=1.0 if args.use_distill else 4.0,
+                guidance_scale_2=1.0 if args.use_distill else 4.0,
+                num_inference_steps=args.num_inference_steps,
+                num_frames=args.num_frames,
+                fps=args.fps,
+                extra_args={
+                    "stage": args.stage,
+                    "resolution": args.resolution,
+                    "save_fps": args.fps,
+                    "use_distill": args.use_distill,
+                    "use_int8": args.use_int8,
+                    "input_json": args.input_json,
+                    "num_segments": args.num_segments,
+                    "num_cond_frames": args.num_cond_frames,
+                    "ref_img_index": args.ref_img_index,
+                    "mask_frame_range": args.mask_frame_range,
+                    "use_kv_cache": args.use_kv_cache,
+                    "offload_kv_cache": args.offload_kv_cache,
+                },
+            ),
+            use_tqdm=True,
+        )
+    finally:
+        omni.close()
+
+    with tempfile.TemporaryDirectory(prefix="longcat_avatar_audio_mux_") as tmp:
+        if input_case is not None and isinstance(audio_input, dict):
+            audio_for_mux = _create_merged_audio_for_case(
+                audio_input,
+                input_case.get("audio_type", "para"),
+                Path(tmp) / "merged_audio.wav",
+            )
+        _save_video_with_audio(_extract_frames(output), Path(args.output), audio_for_mux, args.fps)
+    print(f"Saved video to {args.output}")
+
+
+if __name__ == "__main__":
+    main()
@@ -43,6 +43,23 @@ This folder provides several entrypoints for experimenting with text-to-image di

 Default model: `Qwen/Qwen-Image`

+### LingBot-Video
+
+LingBot-Video uses the same checkpoint for image and video generation. Run the
+shared T2I entry point with:
+
+```bash
+python examples/offline_inference/text_to_image/text_to_image.py \
+  --model robbyant/lingbot-video-dense-1.3b \
+  --prompt "a red fox standing in fresh snow" \
+  --height 192 --width 320 --num-inference-steps 2 \
+  --guidance-scale 3.0 --extra-body '{"flow_shift":3.0}' \
+  --output lingbot_t2i.png
+```
+
+The LingBot prompt builder selects the `image` output modality and requests one
+frame. Passing an input image is not supported in LingBot T2I mode.
+
 ## Quick Start

 ### Python API
@@ -92,8 +109,9 @@ python text_to_image.py \
 | `--ring-degree` | int | `1` | Ring sequence parallel degree for hybrid Ulysses + Ring inference |
 | `--ulysses-mode` | str | `"strict"` | Ulysses SP mode: `"strict"` or `"advanced_uaa"` |
 | `--enable-cpu-offload` | flag | off | Enable CPU offloading for diffusion models |
-| `--lora-path` | str | — | Path to PEFT LoRA adapter folder |
+| `--lora-path` | str | — | Path to PEFT LoRA adapter folder or checkpoint file |
 | `--lora-scale` | float | `1.0` | Scale factor for LoRA weights |
+| `--lora-backend` | str |`"peft"`| LoRA backend for loading LoRA adapters. Default: peft. Available options: peft, distill |
 | `--use-system-prompt` | str | `None` | System prompt preset: `en_unified`, `en_vanilla`, `en_recaption`, `en_think_recaption`, `dynamic`, `None`, or custom text. Recommended: `en_unified`. Only for HunyuanImage-3.0.|
 | `--system-prompt` | str | `None` | Custom system prompt text. Only used when `--use-system-prompt` is set to `custom`. Only for HunyuanImage-3.0.|
 | `--auxiliary-text-encoder` | str | `None` | Supplementary auxiliary text encoder parameters model name or path (especially for Hidream-l1-full). |
@@ -8,7 +8,9 @@ import time
 from pathlib import Path
 from typing import Any

+import numpy as np
 import torch
+from diffusers.utils import numpy_to_pil

 from vllm_omni.diffusion.data import logger
 from vllm_omni.diffusion.utils.image_output import extract_images_from_outputs
@@ -54,6 +56,17 @@ def parse_json_object(value: str, flag_name: str = "argument") -> dict[str, Any]
 parse_profiler_config = functools.partial(parse_json_object, flag_name="--profiler-config")


+def _normalize_images_for_save(images: list[Any]) -> list[Any]:
+    """Convert NumPy diffusion outputs to PIL images before saving."""
+    normalized = []
+    for image in images:
+        if isinstance(image, np.ndarray):
+            normalized.extend(numpy_to_pil(image))
+        else:
+            normalized.append(image)
+    return normalized
+
+
 def parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description="Generate an image with supported diffusion models.")
    parser.add_argument(
@@ -247,6 +260,7 @@ def parse_args() -> argparse.Namespace:
    parser.add_argument(
        "--lora-path",
        type=str,
+        nargs="+",
        default=None,
        help="Path to LoRA adapter folder (PEFT format). Loaded at initialization and used for generation.",
    )
@@ -256,6 +270,15 @@ def parse_args() -> argparse.Namespace:
        default=1.0,
        help="Scale factor for LoRA weights (default: 1.0).",
    )
+    parser.add_argument(
+        "--lora-backend",
+        type=str,
+        default="peft",
+        choices=["peft", "distill"],
+        help="LoRA backend for loading LoRA adapters. Default: peft"
+        "'peft' loads a PEFT-format adapter folder, used e.g. for RL"
+        "'distill' fuses one or more concrete LoRA checkpoint files, used e.g. for distilled few-step LoRAs",
+    )
    parser.add_argument(
        "--vae-patch-parallel-size",
        type=int,
@@ -391,8 +414,12 @@ def main():
    # Prepare LoRA kwargs for Omni initialization
    lora_args: dict[str, Any] = {}
    if args.lora_path:
-        lora_args["lora_path"] = args.lora_path
-        print(f"Using LoRA from: {args.lora_path}")
+        lora_path = args.lora_path
+        if len(lora_path) == 1:
+            lora_path = lora_path[0]
+        lora_args["lora_path"] = lora_path
+        lora_args["lora_backend"] = args.lora_backend
+        print(f"Using LoRA from: {lora_path} with backend: {args.lora_backend}")

    # Build quantization kwargs: use quantization_config dict when
    # ignored_layers is specified so the list flows through OmniDiffusionConfig
@@ -481,12 +508,16 @@ def main():

    # Build LoRA request when --lora-path is set
    lora_request = None
-    if args.lora_path:
-        lora_request_id = stable_lora_int_id(args.lora_path)
+    if args.lora_path and args.lora_backend == "peft":
+        if len(args.lora_path) != 1:
+            raise ValueError("Only one LoRA path is expected for PEFT backend.")
+
+        lora_path = args.lora_path[0]
+        lora_request_id = stable_lora_int_id(lora_path)
        lora_request = LoRARequest(
-            lora_name=Path(args.lora_path).stem,
+            lora_name=Path(lora_path).stem,
            lora_int_id=lora_request_id,
-            lora_path=args.lora_path,
+            lora_path=lora_path,
        )

    generation_start = time.perf_counter()
@@ -619,6 +650,7 @@ def main():

    if not images:
        raise ValueError("No images found in request_output")
+    images = _normalize_images_for_save(images)

    output_path = Path(args.output)
    output_path.parent.mkdir(parents=True, exist_ok=True)
@@ -55,6 +55,21 @@ python text_to_video.py \
  --output vace_t2v_output.mp4
 ```

+### LingBot-Video
+
+The dedicated LingBot example explicitly selects the video output modality:
+
+```bash
+python examples/offline_inference/text_to_video/text_to_video_lingbot.py \
+  --model robbyant/lingbot-video-dense-1.3b \
+  --height 192 --width 320 --num-frames 9 --num-inference-steps 2 \
+  --output lingbot_t2v.mp4
+```
+
+Use the same command with `robbyant/lingbot-video-moe-30b-a3b` for the MoE
+checkpoint. Requested frame counts are rounded upward to the causal
+VAE's `4n+1` grid.
+
 ### LTX-2

 ```bash
@@ -203,6 +218,9 @@ python text_to_video.py \
 - `--audio-sample-rate`: fallback audio sample rate when the pipeline returns audio.
 - `--quantization`: quantization method (such as `fp8` for FP8).
 - `--flow-shift`: scheduler flow_shift parameter.
+- `--lora-path`: path to PEFT LoRA adapter folder or checkpoint file.
+- `--lora-scale`: scale factor for LoRA weights.
+- `--lora-backend`: backend for loading LoRA adapters. Default: peft. Available options: peft, distill.
 - `--extra-body`: JSON object of model-specific generation params, filtered against the model's declared `extra_body_params` (see [`vllm_omni/model_extras`](../../../vllm_omni/model_extras)). Used by Cosmos3 (see above).

 ### Wan2.2-specific
@@ -14,6 +14,8 @@ 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.lora.request import LoRARequest
+from vllm_omni.lora.utils import stable_lora_int_id
 from vllm_omni.model_extras import get_extra_body_params, get_model_class_name
 from vllm_omni.outputs import OmniRequestOutput
 from vllm_omni.platforms import current_omni_platform
@@ -334,6 +336,29 @@ def parse_args() -> argparse.Namespace:
        action="store_true",
        help="Enable expert parallelism for MoE layers.",
    )
+    parser.add_argument(
+        "--lora-path",
+        type=str,
+        nargs="+",
+        default=None,
+        help="Path to LoRA adapter folder (PEFT format) or concrete LoRA checkpoint files. Loaded at initialization and used for generation."
+        "Note: for Wan2.2 MoE models, two checkpoints positionally mapped to high-noise and low-noise modules.",
+    )
+    parser.add_argument(
+        "--lora-scale",
+        type=float,
+        default=1.0,
+        help="Scale factor for LoRA weights (default: 1.0).",
+    )
+    parser.add_argument(
+        "--lora-backend",
+        type=str,
+        default="peft",
+        choices=["peft", "distill"],
+        help="LoRA backend for loading LoRA adapters. Default: peft"
+        "'peft' loads a PEFT-format adapter folder, used e.g. for RL"
+        "'distill' fuses one or more concrete LoRA checkpoint files, used e.g. for distilled few-step LoRAs",
+    )
    parser.add_argument(
        "--use-hsdp",
        action="store_true",
@@ -440,6 +465,12 @@ def main():
        omni_kwargs["cache_backend"] = args.cache_backend
        omni_kwargs["cache_config"] = cache_config
        omni_kwargs["enable_cache_dit_summary"] = args.enable_cache_dit_summary
+    if args.lora_path is not None:
+        lora_path = args.lora_path
+        if len(lora_path) == 1:
+            lora_path = lora_path[0]
+        omni_kwargs["lora_path"] = lora_path
+        omni_kwargs["lora_backend"] = args.lora_backend

    # Cosmos3 loads its (gated) guardrail models at build time, so the guardrails
    # gate is an engine-level config (offline analog of the server's --no-guardrails).
@@ -469,6 +500,19 @@ def main():
    print(f"  Video size: {args.width}x{args.height}")
    print(f"{'=' * 60}\n")

+    lora_request = None
+    if args.lora_path and args.lora_backend == "peft":
+        if len(args.lora_path) != 1:
+            raise ValueError("Only one LoRA path is expected for PEFT backend.")
+
+        lora_path = args.lora_path[0]
+        lora_request_id = stable_lora_int_id(lora_path)
+        lora_request = LoRARequest(
+            lora_name=Path(lora_path).stem,
+            lora_int_id=lora_request_id,
+            lora_path=lora_path,
+        )
+
    prompt_dict = {"prompt": args.prompt}
    if args.negative_prompt is not None:
        prompt_dict["negative_prompt"] = args.negative_prompt
@@ -476,6 +520,11 @@ def main():
        # Preserve the historical empty-prompt behavior for non-LTX examples.
        prompt_dict["negative_prompt"] = ""

+    extra_args = {}
+    if lora_request:
+        extra_args["lora_request"] = lora_request
+        extra_args["lora_scale"] = args.lora_scale
+
    sampling_kwargs = dict(
        height=args.height,
        width=args.width,
@@ -483,6 +532,7 @@ def main():
        guidance_scale=args.guidance_scale,
        num_inference_steps=args.num_inference_steps,
        num_frames=args.num_frames,
+        extra_args=extra_args,
    )
    if args.guidance_scale_high is not None:
        sampling_kwargs["guidance_scale_2"] = args.guidance_scale_high
@@ -20,9 +20,9 @@ from vllm_omni.platforms import current_omni_platform

 def parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description="Generate text-to-video with a LingBot-Video checkpoint.")
-    parser.add_argument("--model", default="/home/models/lingbot-video-dense-1.3b")
+    parser.add_argument("--model", default="robbyant/lingbot-video-dense-1.3b")
    parser.add_argument("--prompt", default="a robotic arm picks up a red block")
-    parser.add_argument("--negative-prompt", default="")
+    parser.add_argument("--negative-prompt", default=None)
    parser.add_argument("--output", default="lingbot_video_output.mp4")
    parser.add_argument("--height", type=int, default=192)
    parser.add_argument("--width", type=int, default=320)
@@ -58,9 +58,12 @@ def _to_frame_array(video: Any) -> np.ndarray:
    if isinstance(video, torch.Tensor):
        tensor = video.detach().cpu()
        if tensor.dim() == 5:
-            tensor = tensor[0] if tensor.shape[1] not in (3, 4) else tensor[0].permute(1, 2, 3, 0)
-        elif tensor.dim() == 4 and tensor.shape[0] in (3, 4):
-            tensor = tensor.permute(1, 2, 3, 0)
+            tensor = tensor[0]
+        if tensor.dim() == 4 and tensor.shape[-1] not in (3, 4):
+            if tensor.shape[0] in (3, 4):
+                tensor = tensor.permute(1, 2, 3, 0)
+            elif tensor.shape[1] in (3, 4):
+                tensor = tensor.permute(0, 2, 3, 1)
        if tensor.is_floating_point():
            if float(tensor.min()) < 0.0:
                tensor = tensor.clamp(-1, 1) * 0.5 + 0.5
@@ -87,8 +90,12 @@ def main() -> None:
        parallel_config=DiffusionParallelConfig(),
    )

-    prompt: dict[str, str] = {"prompt": args.prompt}
-    if args.negative_prompt:
+    prompt: dict[str, Any] = {
+        "prompt": args.prompt,
+        "modalities": ["video"],
+        "num_frames": args.num_frames,
+    }
+    if args.negative_prompt is not None:
        prompt["negative_prompt"] = args.negative_prompt

    sampling_params = OmniDiffusionSamplingParams(
@@ -0,0 +1,74 @@
+# LingBot-Video
+
+LingBot-Video uses one `LingBotVideoPipeline` for text-to-image (T2I),
+text-to-video (T2V), and text-image-to-video (TI2V) generation. Both the dense
+and MoE checkpoints use the same request format.
+
+## Start the server
+
+```bash
+MODEL=robbyant/lingbot-video-dense-1.3b bash run_server.sh
+```
+
+The MoE checkpoint uses the same server and request scripts, but requires
+substantially more GPU memory:
+
+```bash
+MODEL=robbyant/lingbot-video-moe-30b-a3b bash run_server.sh
+```
+
+## Text to image
+
+The image endpoint selects T2I mode and always generates one frame:
+
+```bash
+bash run_curl_text_to_image.sh
+```
+
+The script sends a `320x192`, two-step smoke request and writes
+`lingbot_t2i.png`.
+
+## Text or text-image to video
+
+Run the video script without an image to select T2V mode:
+
+```bash
+bash run_curl_text_image_to_video.sh
+```
+
+Pass a first-frame image to the same script to select TI2V mode:
+
+```bash
+INPUT_IMAGE=/path/to/input.png bash run_curl_text_image_to_video.sh
+```
+
+The client scripts omit the optional `model` request field, so they target
+whichever dense or MoE checkpoint the server loaded. The video example uses the
+lightweight `320x192`, 9-frame, two-step configuration.
+
+Until the shared `/v1/videos` reference-image resizing is removed, TI2V target
+dimensions must be sent through `extra_params`, for example
+`{"size":"320x192"}`. Do not use the top-level `size`, `width`, or `height`
+fields for TI2V because the serving layer currently applies those dimensions
+to the reference image before the model receives it. T2V requests can continue
+to use the top-level dimension fields.
+
+LingBot video frame counts use the causal VAE `4n+1` grid. The pipeline rounds
+any requested frame count upward to the next valid value. An explicit
+`num_frames` takes precedence over `seconds`; otherwise, the server first
+resolves `seconds * fps` and the pipeline applies the same alignment.
+
+Official `resolution`/`ratio` presets can be sent through `extra_params`, for
+example `{"resolution":"720p","ratio":"16:9"}`. The `2k` and `4k` entries
+only define output dimensions; whether they run successfully depends on the
+checkpoint, GPU memory, and memory optimizations available in the deployment.
+
+For `/v1/images/generations`, the server resolves these aliases to their final
+pixel dimensions before applying `--max-generated-image-size`. Requests above
+the configured limit return HTTP 400 before engine dispatch. LingBot produces
+one output per prompt; image requests with `n>1` are also rejected with HTTP
+400.
+
+LingBot TI2V accepts exactly one image reference. Image editing, video
+references, audio references, batching, and Refiner execution are not supported
+by this pipeline mode.
@@ -0,0 +1,67 @@
+#!/bin/bash
+
+set -euo pipefail
+
+if [ -n "${INPUT_IMAGE:-}" ] && [ ! -f "${INPUT_IMAGE}" ]; then
+    echo "INPUT_IMAGE does not exist: ${INPUT_IMAGE}"
+    exit 1
+fi
+
+BASE_URL="${BASE_URL:-http://localhost:8099}"
+POLL_INTERVAL="${POLL_INTERVAL:-2}"
+
+if [ -n "${INPUT_IMAGE:-}" ]; then
+    default_output="lingbot_ti2v.mp4"
+else
+    default_output="lingbot_t2v.mp4"
+fi
+OUTPUT_PATH="${OUTPUT_PATH:-${default_output}}"
+
+curl_args=(
+    -sS -X POST "${BASE_URL}/v1/videos"
+    -H "Accept: application/json"
+    -F "prompt=the subject turns toward the camera with smooth natural motion"
+    -F "num_frames=9"
+    -F "fps=24"
+    -F "num_inference_steps=2"
+    -F "guidance_scale=3.0"
+    -F "flow_shift=3.0"
+    -F "seed=42"
+)
+if [ -n "${INPUT_IMAGE:-}" ]; then
+    curl_args+=(
+        -F "input_reference=@${INPUT_IMAGE}"
+        -F 'extra_params={"size":"320x192"}'
+    )
+else
+    curl_args+=(-F "width=320" -F "height=192")
+fi
+
+create_response="$(curl "${curl_args[@]}")"
+
+video_id="$(echo "${create_response}" | jq -er '.id')"
+echo "Created video job ${video_id}"
+
+while true; do
+    status_response="$(curl -sS "${BASE_URL}/v1/videos/${video_id}")"
+    status="$(echo "${status_response}" | jq -er '.status')"
+    case "${status}" in
+        queued|in_progress)
+            sleep "${POLL_INTERVAL}"
+            ;;
+        completed)
+            break
+            ;;
+        failed)
+            echo "${status_response}" | jq .
+            exit 1
+            ;;
+        *)
+            echo "Unexpected video job status: ${status}"
+            exit 1
+            ;;
+    esac
+done
+
+curl -sS -L "${BASE_URL}/v1/videos/${video_id}/content" -o "${OUTPUT_PATH}"
+echo "Saved video to ${OUTPUT_PATH}"
@@ -0,0 +1,20 @@
+#!/bin/bash
+
+set -euo pipefail
+
+BASE_URL="${BASE_URL:-http://localhost:8099}"
+OUTPUT_PATH="${OUTPUT_PATH:-lingbot_t2i.png}"
+
+curl -sS -X POST "${BASE_URL}/v1/images/generations" \
+    -H "Content-Type: application/json" \
+    -d "{
+        \"prompt\": \"a red fox standing in fresh snow\",
+        \"size\": \"320x192\",
+        \"num_inference_steps\": 2,
+        \"guidance_scale\": 3.0,
+        \"flow_shift\": 3.0,
+        \"seed\": 42,
+        \"response_format\": \"b64_json\"
+    }" | jq -er '.data[0].b64_json' | base64 -d > "${OUTPUT_PATH}"
+
+echo "Saved image to ${OUTPUT_PATH}"
@@ -0,0 +1,10 @@
+#!/bin/bash
+
+set -euo pipefail
+
+MODEL="${MODEL:-robbyant/lingbot-video-dense-1.3b}"
+PORT="${PORT:-8099}"
+
+vllm serve "${MODEL}" --omni \
+    --model-class-name LingBotVideoPipeline \
+    --port "${PORT}"
@@ -23,10 +23,10 @@ including `librosa`.

 The deploy config auto-loads via `--omni`.
 The default `vllm_omni/deploy/minicpmo_4_5.yaml` keeps all three stages on
-logical device 0 with memory budgets of 55%, 15%, and 15%, leaving headroom
+logical device 0 with memory budgets of 55%, 15%, and 18%, leaving headroom
 for runtime kernels such as the HiFi-GAN vocoder's cuDNN workspace. The
-profile admits at most four sequences per stage and bounds startup video
-profiling to 32 frames per video. For throughput,
+profile admits at most four sequences per stage and bounds Talker context
+to 4096 tokens. For throughput,
 `minicpmo_4_5_2gpu.yaml` gives the Thinker
 GPU 0 (90%) and colocates the Talker (55%) and Code2Wav (35%) on GPU 1. That
 profile admits at most four concurrent sequences per stage.
@@ -196,6 +196,7 @@ async def run_demo(args: argparse.Namespace) -> dict[str, object]:
            args.model,
            ref_audio=_ref_audio_data_url(args.ref_audio),
            session_id=args.session_id,
+            temperature=args.temperature,
            timeout_s=args.timeout_s,
        )
        stream_event_cursor = len(client.events.events)
@@ -383,6 +384,12 @@ def parse_args() -> argparse.Namespace:
    parser.add_argument("--output-dir", default="/tmp/minicpmo_realtime_duplex_demo")
    parser.add_argument("--chunk-ms", type=int, default=200)
    parser.add_argument("--timeout-s", type=float, default=60.0)
+    parser.add_argument(
+        "--temperature",
+        type=float,
+        default=None,
+        help="Stage0 sampling temperature; response-required soft-interrupt uses 0.0.",
+    )
    parser.add_argument("--no-realtime-pacing", action="store_true")
    parser.add_argument("--require-audio", action="store_true")
    return parser.parse_args()
@@ -115,7 +115,23 @@ python gradio_demo.py

 ## LoRA

-This example supports Peft-compatible LoRA (Low-Rank Adaptation) adapters for diffusion models. The LoRA adapter path must be readable on the **server** machine (usually a local path or a mounted directory).
+This example supports both startup-time distilled LoRAs and request-time
+PEFT-compatible adapters. The LoRA path must be readable on the **server**
+machine (usually a local path or a mounted directory).
+
+### Start Server with a Distilled LoRA
+
+```bash
+vllm serve Qwen/Qwen-Image-2512 \
+  --omni \
+  --port 8091 \
+  --lora-backend distill \
+  --lora-path /path/to/Qwen-Image-2512-Lightning-4steps.safetensors
+```
+
+The distilled LoRA is fused once during server initialization and applies to
+every request. Use the checkpoint's few-step sampling settings, for example
+`num_inference_steps=4` and `true_cfg_scale=1.0`.

 ### Using Python Client with LoRA

@@ -1,247 +0,0 @@
-#!/usr/bin/env python3
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""Verify a ModelOpt FP8 diffusers checkpoint exported by
-quantize_wan2_2_modelopt_fp8.py / quantize_hunyuanvideo_15_modelopt_fp8.py.
-
-Three checks:
-  A. transformer/config.json has a sane quantization_config block.
-  B. transformer/*.safetensors contains FP8 (float8_e4m3fn) quantized tensors.
-  C. transformer disk size is materially smaller than a BF16 baseline.
-
-Example:
-    python examples/quantization/check_modelopt_fp8_export.py \\
-        --output ./hv15-480p-modelopt-fp8
-
-    # Optional: compare disk size against a local or HF BF16 baseline.
-    python examples/quantization/check_modelopt_fp8_export.py \\
-        --output ./hv15-480p-modelopt-fp8 \\
-        --baseline hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v
-"""
-
-from __future__ import annotations
-
-import argparse
-import json
-import sys
-from collections import Counter
-from pathlib import Path
-
-SUPPORTED_ALGOS = ("FP8",)
-
-
-def _check_config(transformer_dir: Path) -> tuple[int, str | None]:
-    """Returns (status, quant_algo). status: 0 pass, 1 fail, 2 warn."""
-    cfg_path = transformer_dir / "config.json"
-    if not cfg_path.exists():
-        print(f"[FAIL] {cfg_path} missing.")
-        return 1, None
-
-    with cfg_path.open(encoding="utf-8") as f:
-        cfg = json.load(f)
-
-    qc = cfg.get("quantization_config")
-    if not isinstance(qc, dict):
-        print(f"[FAIL] No `quantization_config` block in {cfg_path}.")
-        return 1, None
-
-    print(f"[A] quantization_config from {cfg_path}:")
-    print(json.dumps(qc, indent=2))
-
-    quant_algo = qc.get("quant_algo")
-    issues = []
-    if qc.get("quant_method") != "modelopt":
-        issues.append(f"quant_method={qc.get('quant_method')!r} (expected 'modelopt')")
-    if quant_algo not in SUPPORTED_ALGOS:
-        issues.append(
-            f"quant_algo={quant_algo!r} (expected one of {SUPPORTED_ALGOS} — vllm-omni adapter may not auto-detect)"
-        )
-
-    if issues:
-        print("[A] WARN — config looks incomplete:")
-        for issue in issues:
-            print(f"    - {issue}")
-        return 2, quant_algo
-    print(f"[A] PASS — config looks correct (quant_algo={quant_algo}).")
-    return 0, quant_algo
-
-
-def _read_safetensors_header(path: Path) -> dict:
-    """Read the JSON header of a safetensors file. Bypass-safe — doesn't materialize tensors.
-
-    Returns {tensor_name: {'dtype': 'F8_E4M3', 'shape': [...], 'data_offsets': [...]}}.
-    Header dtype strings: F8_E4M3, F8_E5M2, BF16, F16, F32, F64, I8, I16, I32, I64, BOOL, U8, ...
-    """
-    import struct
-
-    with open(path, "rb") as f:
-        header_len = struct.unpack("<Q", f.read(8))[0]
-        header = json.loads(f.read(header_len))
-    header.pop("__metadata__", None)
-    return header
-
-
-def _classify_weight_scale_granularity(weight_scale_shapes: list[list[int]]) -> str:
-    """Infer per-tensor vs per-channel vs per-block from sample weight_scale shapes.
-
-    ModelOpt block-wise produces shapes like `[16, 1, 16, 1]` (broadcasting dims of 1
-    interleaved with block-count dims). We count "meaningful" dims — ones with size > 1 —
-    and classify: 0 meaningful dims = per-tensor (scalar), 1 = per-channel, 2+ = per-block.
-    """
-    if not weight_scale_shapes:
-        return "no weight_scale tensors found"
-
-    def meaningful_dims(shape: list[int]) -> int:
-        return sum(1 for d in shape if d > 1)
-
-    per_tensor = sum(1 for s in weight_scale_shapes if meaningful_dims(s) == 0)
-    per_channel = sum(1 for s in weight_scale_shapes if meaningful_dims(s) == 1)
-    per_block = sum(1 for s in weight_scale_shapes if meaningful_dims(s) >= 2)
-    total = len(weight_scale_shapes)
-    if per_tensor == total:
-        return "per-tensor (all scalar scales)"
-    if per_channel == total:
-        return "per-channel (1 meaningful dim)"
-    if per_block == total:
-        return "per-block (2+ meaningful dims — e.g. [M//bm, 1, N//bn, 1] for tiles)"
-    return f"mixed: per-tensor={per_tensor}, per-channel={per_channel}, per-block={per_block} of {total}"
-
-
-def _check_safetensors(transformer_dir: Path) -> int:
-    """Returns 0 on pass, 1 on fail. Reads on-disk dtype from the safetensors header."""
-    files = sorted(transformer_dir.glob("*.safetensors"))
-    if not files:
-        print(f"[FAIL] No *.safetensors in {transformer_dir}.")
-        return 1
-
-    header_dtype_counts: Counter[str] = Counter()
-    sample_quant_weight_keys: list[str] = []
-    sample_scale_keys: list[str] = []
-    weight_scale_shapes: list[list[int]] = []
-    sample_weight_scale_entries: list[tuple[str, list[int]]] = []
-
-    for f in files:
-        try:
-            header = _read_safetensors_header(f)
-        except Exception as exc:
-            print(f"[B] WARN — could not parse header of {f}: {exc}")
-            continue
-        for k, info in header.items():
-            dtype = info.get("dtype", "?")
-            header_dtype_counts[dtype] += 1
-            # FP8 stores weights as F8_E4M3 directly.
-            if dtype.startswith("F8") and k.endswith(".weight") and len(sample_quant_weight_keys) < 5:
-                sample_quant_weight_keys.append(k)
-            if k.endswith(("_scale", ".weight_scale", ".input_scale", "_scale_inv")) and len(sample_scale_keys) < 5:
-                sample_scale_keys.append(k)
-            if k.endswith(".weight_scale"):
-                weight_scale_shapes.append(info.get("shape", []))
-                if len(sample_weight_scale_entries) < 5:
-                    sample_weight_scale_entries.append((k, info.get("shape", [])))
-
-    print(f"\n[B] On-disk dtype counts across {len(files)} safetensors file(s) (from header, not get_tensor):")
-    for dtype, count in sorted(header_dtype_counts.items(), key=lambda kv: -kv[1]):
-        marker = "  <-- FP8" if dtype.startswith("F8") else ""
-        print(f"    {dtype:10s} {count:>6d}{marker}")
-
-    quant_count = sum(c for d, c in header_dtype_counts.items() if d.startswith("F8"))
-    if quant_count == 0:
-        print("[B] FAIL — no FP8 tensors on disk. Calibration likely did not actually quantize the weights.")
-        return 1
-
-    print(f"[B] PASS — {quant_count} FP8 tensors stored on disk.")
-    if sample_quant_weight_keys:
-        print(f"    sample quantized weight tensors: {sample_quant_weight_keys[:3]}")
-    if sample_scale_keys:
-        print(f"    sample scale tensors:            {sample_scale_keys[:3]}")
-    print("    (Note: torch's get_tensor() may return these as bf16 views on some versions —")
-    print("     irrelevant; vLLM's loader uses native FP8 ops.)")
-
-    # Weight-scale granularity — per-tensor (scalar) vs per-channel (1-D) vs per-block (N-D).
-    print(f"\n    weight_scale granularity: {_classify_weight_scale_granularity(weight_scale_shapes)}")
-    for key, shape in sample_weight_scale_entries[:3]:
-        print(f"      {key}: shape {shape}")
-    return 0
-
-
-def _disk_size_gib(p: Path) -> float:
-    return sum(f.stat().st_size for f in p.rglob("*") if f.is_file()) / (1024**3)
-
-
-def _check_size_vs_baseline(transformer_dir: Path, baseline: str | None) -> int:
-    """Returns 0 always (informational only)."""
-    quant_size = _disk_size_gib(transformer_dir)
-    print(f"\n[C] FP8 transformer disk size: {quant_size:.2f} GiB")
-
-    if baseline is None:
-        print("[C] SKIP — pass --baseline <path or HF id> to compare against BF16.")
-        return 0
-
-    baseline_path = Path(baseline)
-    if not baseline_path.exists():
-        # Try HF download.
-        try:
-            from huggingface_hub import snapshot_download
-        except ImportError:
-            print("[C] SKIP — huggingface_hub not installed and baseline not a local path.")
-            return 0
-        print(f"    Downloading baseline transformer from HF: {baseline}")
-        baseline_path = Path(snapshot_download(baseline, allow_patterns=["transformer/*"]))
-
-    bf16_dir = baseline_path / "transformer" if (baseline_path / "transformer").exists() else baseline_path
-    bf16_size = _disk_size_gib(bf16_dir)
-    if bf16_size == 0:
-        print(f"[C] WARN — baseline transformer dir empty: {bf16_dir}")
-        return 0
-
-    # Expected reduction for FP8: ~50%.
-    min_reduction = 30
-    reduction = (1 - quant_size / bf16_size) * 100
-    print(f"[C] BF16 baseline transformer disk size: {bf16_size:.2f} GiB ({bf16_dir})")
-    print(f"[C] Disk reduction: {reduction:.1f}%  (FP8 transformer is {quant_size / bf16_size:.0%} of BF16)")
-    if reduction < min_reduction:
-        print(
-            f"[C] WARN — FP8 should typically reduce disk by ~40-50%; <{min_reduction}% suggests partial quantization."
-        )
-    return 0
-
-
-def main() -> None:
-    p = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
-    p.add_argument("--output", required=True, help="Path to the exported ModelOpt FP8 checkpoint root.")
-    p.add_argument(
-        "--baseline",
-        default=None,
-        help="Optional BF16 baseline (local diffusers dir or HF id) for disk-size comparison.",
-    )
-    args = p.parse_args()
-
-    out_root = Path(args.output).expanduser().resolve()
-    transformer_dir = out_root / "transformer"
-    if not transformer_dir.exists():
-        print(f"[FAIL] {transformer_dir} does not exist.")
-        sys.exit(1)
-
-    print(f"Checking: {out_root}\n")
-
-    fail = 0
-    config_status, _quant_algo = _check_config(transformer_dir)
-    fail |= config_status
-    fail |= _check_safetensors(transformer_dir)
-    _check_size_vs_baseline(transformer_dir, args.baseline)
-
-    print()
-    if fail == 0:
-        print("=" * 60)
-        print("ALL CHECKS PASSED — checkpoint looks ready for vllm-omni serving.")
-    elif fail == 1:
-        print("=" * 60)
-        print("FAILURES detected — calibration may need to be re-run.")
-        sys.exit(1)
-    else:
-        print("=" * 60)
-        print("WARNINGS only — checkpoint may serve but with caveats. See [A] above.")
-
-
-if __name__ == "__main__":
-    main()
@@ -1,448 +0,0 @@
-#!/usr/bin/env python3
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""Quantize HunyuanVideo-1.5 (480p T2V) to a ModelOpt FP8 Hugging Face checkpoint.
-
-Calibrates the DiT transformer using a small video prompt set and exports a
-diffusers-style directory whose transformer carries ModelOpt FP8 metadata.
-The exported checkpoint is consumable by vllm-omni's ModelOpt FP8 adapter
-(see vllm_omni/diffusion/model_loader/checkpoint_adapters/modelopt.py).
-
-Layers kept full precision match the #2728 / #2795 pattern: modulation,
-AdaLayerNorm, entry/exit projections, embeddings, the token refiner path,
-and final proj_out. MHA quantizers are off by default; HV-1.5 self-attention
-empirically degrades under FP8 (see #2920 ablation).
-
-Example:
-    python examples/quantization/quantize_hunyuanvideo_15_modelopt_fp8.py \\
-        --model hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v \\
-        --output ./hv15-480p-modelopt-fp8 \\
-        --overwrite
-"""
-
-from __future__ import annotations
-
-import argparse
-import copy
-import json
-import re
-import shutil
-import sys
-from pathlib import Path
-from typing import Any
-
-import torch
-from diffusers import DiffusionPipeline
-
-DEFAULT_PROMPTS = [
-    "A dog running across a field of golden wheat.",
-    "An astronaut riding a horse across the surface of Mars, red dust swirling, cinematic wide shot.",
-    "A hummingbird hovering in front of a vibrant red flower, slow motion, macro shot.",
-    "A crackling campfire at night under a starry sky, sparks rising into the dark.",
-    "An underwater shot of a coral reef with tropical fish swimming by, sun rays piercing the water.",
-    "A close-up of a blooming rose covered in morning dew, soft natural light.",
-    "A peaceful mountain village at dawn, mist rolling over the rooftops, cinematic establishing shot.",
-    "A skateboarder doing a kickflip in an urban plaza, slow motion, golden hour lighting.",
-]
-
-
-def _build_parser() -> argparse.ArgumentParser:
-    p = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
-    p.add_argument("--model", required=True, help="Input HV-1.5 diffusers directory or HF id.")
-    p.add_argument("--output", required=True, help="Output directory for the ModelOpt FP8 checkpoint.")
-    p.add_argument("--dtype", choices=("bfloat16", "float16"), default="bfloat16")
-    p.add_argument("--height", type=int, default=480)
-    p.add_argument("--width", type=int, default=832)
-    p.add_argument(
-        "--num-frames",
-        type=int,
-        default=33,
-        help="Frames per calibration sample. 33 matches the typical short benchmark.",
-    )
-    p.add_argument("--guidance-scale", type=float, default=6.0)
-    p.add_argument(
-        "--calib-steps",
-        type=int,
-        default=10,
-        help="Denoising steps per calibration prompt (10 is enough for amax statistics).",
-    )
-    p.add_argument("--calib-size", type=int, default=8, help="How many prompts to use for calibration.")
-    p.add_argument("--seed", type=int, default=42)
-    p.add_argument(
-        "--prompt",
-        action="append",
-        default=[],
-        help="Custom calibration prompt. Repeat to provide multiple.",
-    )
-    p.add_argument(
-        "--quantize-mha",
-        action="store_true",
-        help="Enable FP8 attention K/V/softmax quantizers. Off by default — empirically degrades HV-1.5 video output.",
-    )
-    p.add_argument(
-        "--weight-block-size",
-        type=str,
-        default=None,
-        help="Per-block weight quantization as 'M,N' (e.g. '128,128' for 128x128 tiles). "
-        "Default: per-tensor (one scale per linear). Block-wise typically gives tighter quality at "
-        "negligible memory cost. Static FP8 is exempt from upstream vLLM's online block-wise gate.",
-    )
-    p.add_argument("--overwrite", action="store_true", help="Replace an existing output directory.")
-    return p
-
-
-def _parse_block_size(spec: str | None) -> list[int] | None:
-    if spec is None:
-        return None
-    parts = [int(x) for x in spec.split(",") if x.strip()]
-    if len(parts) != 2:
-        raise SystemExit(f"--weight-block-size must be 'M,N' (2 ints), got {spec!r}")
-    return parts
-
-
-def _require_modelopt() -> Any:
-    try:
-        import modelopt.torch.quantization as mtq
-    except ModuleNotFoundError as exc:
-        raise SystemExit(
-            "NVIDIA ModelOpt is not installed. Install with:\n"
-            "  pip install 'nvidia-modelopt[all]'\n"
-            f"Original error: {exc}"
-        ) from exc
-    return mtq
-
-
-def _ensure_paths(args: argparse.Namespace) -> tuple[str, Path]:
-    model_path = args.model
-    output_dir = Path(args.output).expanduser().resolve()
-    if output_dir.exists():
-        if not args.overwrite:
-            raise SystemExit(f"Output directory already exists: {output_dir}\nPass --overwrite to replace it.")
-        shutil.rmtree(output_dir)
-    return model_path, output_dir
-
-
-def _select_dtype(name: str) -> torch.dtype:
-    return {"bfloat16": torch.bfloat16, "float16": torch.float16}[name]
-
-
-def _build_prompts(args: argparse.Namespace) -> list[str]:
-    prompts = args.prompt or DEFAULT_PROMPTS
-    if args.calib_size <= 0:
-        raise SystemExit("--calib-size must be positive.")
-    if len(prompts) < args.calib_size:
-        repeats = (args.calib_size + len(prompts) - 1) // len(prompts)
-        prompts = (prompts * repeats)[: args.calib_size]
-    return prompts[: args.calib_size]
-
-
-# Layers to KEEP at full precision (mirror of the #2920 wiring + #2728/#2795 skip pattern).
-# - x_embedder, image_embedder, context_embedder*, time_embed*, cond_type_embed: entry/embedding
-# - norm_out, norm1*.linear, norm1_context*.linear, norm2*, norm2_context*: AdaLayerNorm modulation
-# - proj_out: final output projection
-# - token_refiner*: text-encoder refinement uses diffusers raw nn.Linear
-def _filter_func_hv15(name: str) -> bool:
-    pattern = re.compile(
-        r"(proj_out.*|"
-        r".*(x_embedder|image_embedder|context_embedder|context_embedder_2|"
-        r"time_embed|cond_type_embed|"
-        r"norm_out|norm1\.linear|norm1_context\.linear|norm2|norm2_context|"
-        r"token_refiner).*)"
-    )
-    return pattern.match(name) is not None
-
-
-def _mha_filter_func(name: str) -> bool:
-    pattern = re.compile(
-        r".*(q_bmm_quantizer|k_bmm_quantizer|v_bmm_quantizer|softmax_quantizer|bmm2_output_quantizer).*"
-    )
-    return pattern.match(name) is not None
-
-
-def _disable_known_problematic_quantizers(mtq: Any, backbone: torch.nn.Module, *, quantize_mha: bool) -> None:
-    if not hasattr(mtq, "disable_quantizer"):
-        return
-    mtq.disable_quantizer(backbone, _filter_func_hv15)
-    if not quantize_mha:
-        mtq.disable_quantizer(backbone, _mha_filter_func)
-
-
-def _load_pipeline(model_path: str, dtype: torch.dtype) -> DiffusionPipeline:
-    pipe = DiffusionPipeline.from_pretrained(model_path, torch_dtype=dtype)
-    if hasattr(pipe, "set_progress_bar_config"):
-        pipe.set_progress_bar_config(disable=True)
-    pipe.to("cuda")
-    return pipe
-
-
-def _build_forward_loop(pipe: DiffusionPipeline, args: argparse.Namespace, prompts: list[str]):
-    generator = torch.Generator(device="cuda")
-
-    # Try to set guidance on the pipeline's guider object up front (modern
-    # diffusers HV-1.5 uses a Guider abstraction, not a per-call kwarg). Falls
-    # back silently — calibration uses whatever default the pipeline ships with.
-    guider = getattr(pipe, "guider", None)
-    if guider is not None and hasattr(guider, "guidance_scale"):
-        try:
-            guider.guidance_scale = args.guidance_scale
-        except Exception:
-            pass
-
-    base_kwargs = dict(
-        height=args.height,
-        width=args.width,
-        num_frames=args.num_frames,
-        num_inference_steps=args.calib_steps,
-        output_type="latent",
-    )
-
-    def forward_loop(*_unused_args, **_unused_kwargs) -> None:
-        with torch.inference_mode():
-            for idx, prompt in enumerate(prompts):
-                generator.manual_seed(args.seed + idx)
-                # Try with guidance_scale first; fall back without on TypeError
-                # for pipelines (like HV-1.5) that take CFG via guider config.
-                try:
-                    pipe(prompt=prompt, generator=generator, guidance_scale=args.guidance_scale, **base_kwargs)
-                except TypeError as exc:
-                    if "guidance_scale" not in str(exc):
-                        raise
-                    pipe(prompt=prompt, generator=generator, **base_kwargs)
-
-    return forward_loop
-
-
-def _summarize_export(output_dir: Path) -> None:
-    cfg_path = output_dir / "transformer" / "config.json"
-    if not cfg_path.exists():
-        print(f"[warn] {cfg_path} missing.", file=sys.stderr)
-        return
-    with cfg_path.open(encoding="utf-8") as f:
-        cfg = json.load(f)
-    qc = cfg.get("quantization_config")
-    if not isinstance(qc, dict):
-        print("[warn] No quantization_config in transformer/config.json.", file=sys.stderr)
-        return
-    print("Export summary:")
-    print(f"  quant_method: {qc.get('quant_method')}")
-    print(f"  quant_algo:   {qc.get('quant_algo')}")
-    producer = qc.get("producer")
-    if isinstance(producer, dict):
-        print(f"  producer:     {producer.get('name')} {producer.get('version')}")
-    print(f"  config path:  {cfg_path}")
-
-
-def _force_export_quantized_weights(backbone: torch.nn.Module, dtype: torch.dtype) -> int:
-    """Convert in-memory weights of quantized modules to actual FP8 storage.
-
-    `export_hf_checkpoint` skips this step for unknown model types (HV-1.5 isn't
-    in ModelOpt's recognized-model registry), so we must call the per-weight
-    export helper ourselves. Same workaround as the HunyuanImage-3 calibration
-    helper.
-    """
-    from modelopt.torch.export.quant_utils import (
-        QUANTIZATION_NONE,
-        get_quantization_format,
-        quantizer_attr_names,
-        weight_attr_names,
-    )
-    from modelopt.torch.export.unified_export_hf import _export_quantized_weight
-
-    exported = 0
-    for name, module in backbone.named_modules():
-        try:
-            quantization_format = get_quantization_format(module)
-        except Exception as exc:
-            print(f"[warn] Could not inspect quantization format for {name}: {exc}", file=sys.stderr)
-            continue
-        if quantization_format == QUANTIZATION_NONE:
-            continue
-        for weight_name in weight_attr_names(module):
-            quantizer_attrs = quantizer_attr_names(weight_name)
-            weight_quantizer = getattr(module, quantizer_attrs.weight_quantizer, None)
-            if weight_quantizer is None or not getattr(weight_quantizer, "is_enabled", False):
-                continue
-            _export_quantized_weight(module, dtype, weight_name)
-            exported += 1
-    return exported
-
-
-def _hv15_quant_config_block(weight_block_size: list[int] | None = None) -> dict:
-    """Mirror ModelOpt FP8 metadata expected by vllm-omni's adapter (#2913).
-
-    Same shape as the HunyuanImage-3 author's _hunyuan_quant_config(). When
-    `weight_block_size` is given, advertise block-wise weight quantization in
-    the saved metadata (so consumers know to expect multi-element scale tensors).
-    """
-    weights_cfg: dict = {"dynamic": False, "num_bits": 8, "type": "float"}
-    if weight_block_size is not None:
-        weights_cfg["strategy"] = "block"
-        weights_cfg["block_structure"] = f"{weight_block_size[0]}x{weight_block_size[1]}"
-    return {
-        "config_groups": {
-            "group_0": {
-                "input_activations": {"dynamic": False, "num_bits": 8, "type": "float"},
-                "weights": weights_cfg,
-                "targets": ["Linear"],
-            }
-        },
-        "ignore": [
-            "context_embedder*",
-            "context_embedder_2*",
-            "cond_type_embed*",
-            "image_embedder*",
-            "norm1.linear*",
-            "norm1_context.linear*",
-            "norm2*",
-            "norm2_context*",
-            "norm_out*",
-            "proj_out*",
-            "time_embed*",
-            "token_refiner*",
-            "x_embedder*",
-        ],
-        "producer": {"name": "modelopt"},
-        "quant_algo": "FP8",
-        "quant_method": "modelopt",
-    }
-
-
-def _patch_quant_config(output_dir: Path, weight_block_size: list[int] | None = None) -> None:
-    """Inject quant_algo: FP8 + config_groups into transformer/config.json so
-    vllm-omni's adapter (#2913) recognises the checkpoint as ModelOpt FP8."""
-    cfg_path = output_dir / "transformer" / "config.json"
-    with cfg_path.open(encoding="utf-8") as f:
-        cfg = json.load(f)
-
-    new_qc = _hv15_quant_config_block(weight_block_size=weight_block_size)
-    existing = cfg.get("quantization_config")
-    if isinstance(existing, dict):
-        producer = existing.get("producer")
-        if isinstance(producer, dict):
-            new_qc["producer"] = producer
-
-    cfg["quantization_config"] = new_qc
-    with cfg_path.open("w", encoding="utf-8") as f:
-        json.dump(cfg, f, indent=2)
-
-
-def _save_pipeline_with_fp8_transformer(
-    pipe: DiffusionPipeline,
-    model_path: str,
-    output_dir: Path,
-    max_shard_size: str = "5GB",
-) -> None:
-    """Save the pipeline with the (now FP8) transformer.
-
-    Copies the source directory verbatim except for `transformer/`, then
-    saves the transformer with quantizers hidden so the state dict contains
-    only the FP8 weights + scale tensors.
-    """
-    from modelopt.torch.export.diffusers_utils import hide_quantizers_from_state_dict
-
-    src = Path(model_path)
-    if not src.exists():
-        from huggingface_hub import snapshot_download
-
-        src = Path(snapshot_download(model_path))
-
-    if output_dir.exists():
-        shutil.rmtree(output_dir)
-    shutil.copytree(src, output_dir, ignore=shutil.ignore_patterns("transformer"))
-
-    transformer_out = output_dir / "transformer"
-    # `hide_quantizers_from_state_dict` walks named_modules(); pass the actual
-    # nn.Module (transformer), not the diffusers Pipeline wrapper.
-    with hide_quantizers_from_state_dict(pipe.transformer):
-        pipe.transformer.save_pretrained(
-            str(transformer_out),
-            safe_serialization=True,
-            max_shard_size=max_shard_size,
-        )
-
-
-def main() -> None:
-    args = _build_parser().parse_args()
-    if not torch.cuda.is_available():
-        raise SystemExit("CUDA is required for ModelOpt FP8 quantization.")
-
-    mtq = _require_modelopt()
-    model_path, output_dir = _ensure_paths(args)
-    dtype = _select_dtype(args.dtype)
-    prompts = _build_prompts(args)
-
-    print("Quantization plan:")
-    weight_block_size = _parse_block_size(args.weight_block_size)
-
-    print(f"  input:           {args.model}")
-    print(f"  output:          {output_dir}")
-    print(f"  dtype:           {dtype}")
-    print(f"  height/width:    {args.height}x{args.width}")
-    print(f"  num_frames:      {args.num_frames}")
-    print(f"  calib_size:      {len(prompts)}")
-    print(f"  calib_steps:     {args.calib_steps}")
-    print(f"  quantize_mha:    {args.quantize_mha}")
-    print(
-        f"  weight strategy: {'block-wise ' + str(weight_block_size) if weight_block_size else 'per-tensor (default)'}"
-    )
-
-    pipe = _load_pipeline(model_path, dtype)
-    backbone = pipe.transformer
-
-    quant_config = copy.deepcopy(mtq.FP8_DEFAULT_CFG)
-    if weight_block_size is not None:
-        # Switch from per-tensor (default) to block-wise weight quantization.
-        # ModelOpt's wildcard "*weight_quantizer" matches every linear's weight quantizer.
-        quant_config["quant_cfg"]["*weight_quantizer"] = {
-            "num_bits": (4, 3),  # E4M3 (FP8 weights, same as default)
-            "block_sizes": {-1: weight_block_size[1], -2: weight_block_size[0]},
-        }
-        print(
-            f"  -> overriding weight quantizer with block_sizes={weight_block_size} "
-            f"({weight_block_size[0]}x{weight_block_size[1]} tiles)"
-        )
-
-    forward_loop = _build_forward_loop(pipe, args, prompts)
-    quantized = mtq.quantize(backbone, quant_config, forward_loop)
-    if quantized is not None:
-        pipe.transformer = quantized
-        backbone = quantized
-
-    _disable_known_problematic_quantizers(mtq, backbone, quantize_mha=args.quantize_mha)
-
-    print("\nForcing FP8 weight serialization (HV-1.5 isn't in ModelOpt's recognized-model registry,")
-    print("so we have to call the per-weight export helper ourselves)...")
-    exported = _force_export_quantized_weights(backbone, dtype)
-    print(f"  -> {exported} weights converted to FP8 in memory")
-    if exported == 0:
-        raise SystemExit(
-            "No quantized weights were exported. Calibration may have skipped every layer "
-            "(check the disable_quantizer regex) or `mtq.quantize` did not actually wrap any "
-            "weight quantizers."
-        )
-
-    print("\nSaving pipeline with FP8 transformer...")
-    _save_pipeline_with_fp8_transformer(pipe, model_path, output_dir)
-    _patch_quant_config(output_dir, weight_block_size=weight_block_size)
-    print(f"Saved to: {output_dir}")
-    _summarize_export(output_dir)
-
-    print("\nNext: validate the checkpoint with vllm-omni:")
-    print(
-        "  python examples/offline_inference/text_to_video/text_to_video.py \\\n"
-        f"    --model {output_dir} \\\n"
-        "    --quantization fp8 \\\n"
-        "    --prompt 'A dog running across a field of golden wheat.' \\\n"
-        f"    --height {args.height} --width {args.width} --num-frames {args.num_frames} \\\n"
-        "    --num-inference-steps 30 --guidance-scale 6.0 --seed 42 \\\n"
-        "    --output outputs/hv15_modelopt_fp8.mp4 \\\n"
-        "    --enforce-eager"
-    )
-    print(
-        "\n  (--quantization fp8 is auto-upgraded to ModelOpt FP8 at runtime because the "
-        "checkpoint's config.json has modelopt metadata.)"
-    )
-
-
-if __name__ == "__main__":
-    main()
@@ -1,471 +0,0 @@
-#!/usr/bin/env python3
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""Quantize Wan2.2 (TI2V-5B, 704x1280 T2V) to a ModelOpt FP8 Hugging Face checkpoint.
-
-Calibrates the DiT transformer using a small video prompt set and exports a
-diffusers-style directory whose transformer carries ModelOpt FP8 metadata.
-The exported checkpoint is consumable by vllm-omni's ModelOpt FP8 adapter
-(see vllm_omni/diffusion/model_loader/checkpoint_adapters/modelopt.py).
-
-Layers kept full precision match the #2728 / #2795 pattern: condition embedder
-(time/text/image), patch embedding, modulation (scale_shift_table), final
-norm + proj_out, and sequence-parallel helpers. All attention + FFN linears
-are quantized — static calibration handles the numerics that online FP8
-couldn't (see #2920 ablation).
-
-Default target is `Wan-AI/Wan2.2-TI2V-5B-Diffusers`, the dense 5B variant that
-fits 80GB BF16. The A14B MoE variants need 2+ GPUs and are out of scope here.
-
-Example:
-    python examples/quantization/quantize_wan2_2_modelopt_fp8.py \\
-        --model Wan-AI/Wan2.2-TI2V-5B-Diffusers \\
-        --output ./wan22-ti2v-modelopt-fp8 \\
-        --overwrite
-"""
-
-from __future__ import annotations
-
-import argparse
-import copy
-import json
-import re
-import shutil
-import sys
-from pathlib import Path
-from typing import Any
-
-import torch
-from diffusers import DiffusionPipeline
-
-DEFAULT_PROMPTS = [
-    "A dog running across a field of golden wheat.",
-    "An astronaut riding a horse across the surface of Mars, red dust swirling, cinematic wide shot.",
-    "A hummingbird hovering in front of a vibrant red flower, slow motion, macro shot.",
-    "A crackling campfire at night under a starry sky, sparks rising into the dark.",
-    "An underwater shot of a coral reef with tropical fish swimming by, sun rays piercing the water.",
-    "A close-up of a blooming rose covered in morning dew, soft natural light.",
-    "A peaceful mountain village at dawn, mist rolling over the rooftops, cinematic establishing shot.",
-    "A skateboarder doing a kickflip in an urban plaza, slow motion, golden hour lighting.",
-]
-
-
-def _build_parser() -> argparse.ArgumentParser:
-    p = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
-    p.add_argument("--model", required=True, help="Input Wan2.2 diffusers directory or HF id.")
-    p.add_argument("--output", required=True, help="Output directory for the ModelOpt FP8 checkpoint.")
-    p.add_argument("--dtype", choices=("bfloat16", "float16"), default="bfloat16")
-    p.add_argument("--height", type=int, default=704, help="Calibration video height (Wan2.2 TI2V-5B native: 704).")
-    p.add_argument("--width", type=int, default=1280, help="Calibration video width (Wan2.2 TI2V-5B native: 1280).")
-    p.add_argument(
-        "--num-frames",
-        type=int,
-        default=49,
-        help="Frames per calibration sample. 49 matches the typical short benchmark; "
-        "use 17 to reduce memory pressure during calibration.",
-    )
-    p.add_argument("--guidance-scale", type=float, default=5.0)
-    p.add_argument(
-        "--calib-steps",
-        type=int,
-        default=10,
-        help="Denoising steps per calibration prompt (10 is enough for amax statistics).",
-    )
-    p.add_argument("--calib-size", type=int, default=8, help="How many prompts to use for calibration.")
-    p.add_argument("--seed", type=int, default=42)
-    p.add_argument(
-        "--prompt",
-        action="append",
-        default=[],
-        help="Custom calibration prompt. Repeat to provide multiple.",
-    )
-    p.add_argument(
-        "--quantize-mha",
-        action="store_true",
-        help="Enable FP8 attention K/V/softmax quantizers. Off by default — Wan2.2's long attention "
-        "sequences amplified FP8 drift in the online ablation (see #2920).",
-    )
-    p.add_argument(
-        "--weight-block-size",
-        type=str,
-        default=None,
-        help="Per-block weight quantization as 'M,N' (e.g. '128,128'). Default per-tensor. "
-        "Note: vllm-omni's ModelOpt adapter may not yet dispatch block-wise scales — check #2924 "
-        "for the HV-1.5 investigation status before relying on this.",
-    )
-    p.add_argument("--overwrite", action="store_true", help="Replace an existing output directory.")
-    return p
-
-
-def _parse_block_size(spec: str | None) -> list[int] | None:
-    if spec is None:
-        return None
-    parts = [int(x) for x in spec.split(",") if x.strip()]
-    if len(parts) != 2:
-        raise SystemExit(f"--weight-block-size must be 'M,N' (2 ints), got {spec!r}")
-    return parts
-
-
-def _require_modelopt() -> Any:
-    try:
-        import modelopt.torch.quantization as mtq
-    except ModuleNotFoundError as exc:
-        raise SystemExit(
-            "NVIDIA ModelOpt is not installed. Install with:\n"
-            "  pip install 'nvidia-modelopt[all]'\n"
-            f"Original error: {exc}"
-        ) from exc
-    return mtq
-
-
-def _ensure_paths(args: argparse.Namespace) -> tuple[str, Path]:
-    model_path = args.model
-    output_dir = Path(args.output).expanduser().resolve()
-    if output_dir.exists():
-        if not args.overwrite:
-            raise SystemExit(f"Output directory already exists: {output_dir}\nPass --overwrite to replace it.")
-        shutil.rmtree(output_dir)
-    return model_path, output_dir
-
-
-def _select_dtype(name: str) -> torch.dtype:
-    return {"bfloat16": torch.bfloat16, "float16": torch.float16}[name]
-
-
-def _build_prompts(args: argparse.Namespace) -> list[str]:
-    prompts = args.prompt or DEFAULT_PROMPTS
-    if args.calib_size <= 0:
-        raise SystemExit("--calib-size must be positive.")
-    if len(prompts) < args.calib_size:
-        repeats = (args.calib_size + len(prompts) - 1) // len(prompts)
-        prompts = (prompts * repeats)[: args.calib_size]
-    return prompts[: args.calib_size]
-
-
-# Layers to KEEP at full precision. Wan2.2's module naming:
-# - condition_embedder: time_embedder, time_proj, text_embedder, image_embedder (I2V)
-# - patch_embedding: Conv3dLayer (already not Linear, belt-and-suspenders skip)
-# - scale_shift_table: nn.Parameter modulation (not Linear, but pattern guard)
-# - norm_out: AdaLayerNorm final
-# - proj_out: final nn.Linear
-# - timestep_proj_prepare / output_scale_shift_prepare: SP helpers
-def _filter_func_wan22(name: str) -> bool:
-    pattern = re.compile(
-        r"(proj_out.*|"
-        r".*(condition_embedder|patch_embedding|"
-        r"norm_out|scale_shift_table|"
-        r"timestep_proj_prepare|output_scale_shift_prepare).*)"
-    )
-    return pattern.match(name) is not None
-
-
-def _mha_filter_func(name: str) -> bool:
-    pattern = re.compile(
-        r".*(q_bmm_quantizer|k_bmm_quantizer|v_bmm_quantizer|softmax_quantizer|bmm2_output_quantizer).*"
-    )
-    return pattern.match(name) is not None
-
-
-def _disable_known_problematic_quantizers(mtq: Any, backbone: torch.nn.Module, *, quantize_mha: bool) -> None:
-    if not hasattr(mtq, "disable_quantizer"):
-        return
-    mtq.disable_quantizer(backbone, _filter_func_wan22)
-    if not quantize_mha:
-        mtq.disable_quantizer(backbone, _mha_filter_func)
-
-
-def _load_pipeline(model_path: str, dtype: torch.dtype) -> DiffusionPipeline:
-    pipe = DiffusionPipeline.from_pretrained(model_path, torch_dtype=dtype)
-    if hasattr(pipe, "set_progress_bar_config"):
-        pipe.set_progress_bar_config(disable=True)
-    pipe.to("cuda")
-    return pipe
-
-
-def _build_forward_loop(pipe: DiffusionPipeline, args: argparse.Namespace, prompts: list[str]):
-    generator = torch.Generator(device="cuda")
-
-    # Try setting guidance on the pipeline's guider if present (newer diffusers APIs).
-    guider = getattr(pipe, "guider", None)
-    if guider is not None and hasattr(guider, "guidance_scale"):
-        try:
-            guider.guidance_scale = args.guidance_scale
-        except Exception:
-            pass
-
-    base_kwargs = dict(
-        height=args.height,
-        width=args.width,
-        num_frames=args.num_frames,
-        num_inference_steps=args.calib_steps,
-        output_type="latent",
-    )
-
-    def forward_loop(*_unused_args, **_unused_kwargs) -> None:
-        with torch.inference_mode():
-            for idx, prompt in enumerate(prompts):
-                generator.manual_seed(args.seed + idx)
-                # Try with guidance_scale first; fall back without on TypeError
-                # for pipelines that take CFG via guider config only.
-                try:
-                    pipe(prompt=prompt, generator=generator, guidance_scale=args.guidance_scale, **base_kwargs)
-                except TypeError as exc:
-                    if "guidance_scale" not in str(exc):
-                        raise
-                    pipe(prompt=prompt, generator=generator, **base_kwargs)
-
-    return forward_loop
-
-
-def _summarize_export(output_dir: Path) -> None:
-    cfg_path = output_dir / "transformer" / "config.json"
-    if not cfg_path.exists():
-        print(f"[warn] {cfg_path} missing.", file=sys.stderr)
-        return
-    with cfg_path.open(encoding="utf-8") as f:
-        cfg = json.load(f)
-    qc = cfg.get("quantization_config")
-    if not isinstance(qc, dict):
-        print("[warn] No quantization_config in transformer/config.json.", file=sys.stderr)
-        return
-    print("Export summary:")
-    print(f"  quant_method: {qc.get('quant_method')}")
-    print(f"  quant_algo:   {qc.get('quant_algo')}")
-    producer = qc.get("producer")
-    if isinstance(producer, dict):
-        print(f"  producer:     {producer.get('name')} {producer.get('version')}")
-    print(f"  config path:  {cfg_path}")
-
-
-def _force_export_quantized_weights(backbone: torch.nn.Module, dtype: torch.dtype) -> int:
-    """Convert in-memory weights of quantized modules to actual FP8 storage.
-
-    `export_hf_checkpoint` skips this step for unknown model types (Wan2.2 isn't
-    in ModelOpt's recognized-model registry), so we must call the per-weight
-    export helper ourselves. Same workaround as the HunyuanVideo-1.5 / HunyuanImage-3
-    calibration helpers.
-    """
-    from modelopt.torch.export.quant_utils import (
-        QUANTIZATION_NONE,
-        get_quantization_format,
-        quantizer_attr_names,
-        weight_attr_names,
-    )
-    from modelopt.torch.export.unified_export_hf import _export_quantized_weight
-
-    exported = 0
-    for name, module in backbone.named_modules():
-        try:
-            quantization_format = get_quantization_format(module)
-        except Exception as exc:
-            print(f"[warn] Could not inspect quantization format for {name}: {exc}", file=sys.stderr)
-            continue
-        if quantization_format == QUANTIZATION_NONE:
-            continue
-        for weight_name in weight_attr_names(module):
-            quantizer_attrs = quantizer_attr_names(weight_name)
-            weight_quantizer = getattr(module, quantizer_attrs.weight_quantizer, None)
-            if weight_quantizer is None or not getattr(weight_quantizer, "is_enabled", False):
-                continue
-            _export_quantized_weight(module, dtype, weight_name)
-            exported += 1
-    return exported
-
-
-def _wan22_quant_config_block(weight_block_size: list[int] | None = None) -> dict:
-    """Mirror ModelOpt FP8 metadata expected by vllm-omni's adapter (#2913)."""
-    weights_cfg: dict = {"dynamic": False, "num_bits": 8, "type": "float"}
-    if weight_block_size is not None:
-        weights_cfg["strategy"] = "block"
-        weights_cfg["block_structure"] = f"{weight_block_size[0]}x{weight_block_size[1]}"
-    return {
-        "config_groups": {
-            "group_0": {
-                "input_activations": {"dynamic": False, "num_bits": 8, "type": "float"},
-                "weights": weights_cfg,
-                "targets": ["Linear"],
-            }
-        },
-        "ignore": [
-            "condition_embedder*",
-            "norm_out*",
-            "output_scale_shift_prepare*",
-            "patch_embedding*",
-            "proj_out*",
-            "scale_shift_table*",
-            "timestep_proj_prepare*",
-        ],
-        "producer": {"name": "modelopt"},
-        "quant_algo": "FP8",
-        "quant_method": "modelopt",
-    }
-
-
-def _patch_quant_config(output_dir: Path, weight_block_size: list[int] | None = None) -> None:
-    """Inject quant_algo: FP8 + config_groups into each transformer's config.json
-    so vllm-omni's adapter (#2913) recognises the checkpoint as ModelOpt FP8.
-
-    Patches both ``transformer`` and ``transformer_2`` (A14B) when present."""
-    for sub in ("transformer", "transformer_2"):
-        cfg_path = output_dir / sub / "config.json"
-        if not cfg_path.exists():
-            continue
-        with cfg_path.open(encoding="utf-8") as f:
-            cfg = json.load(f)
-
-        new_qc = _wan22_quant_config_block(weight_block_size=weight_block_size)
-        existing = cfg.get("quantization_config")
-        if isinstance(existing, dict):
-            producer = existing.get("producer")
-            if isinstance(producer, dict):
-                new_qc["producer"] = producer
-
-        cfg["quantization_config"] = new_qc
-        with cfg_path.open("w", encoding="utf-8") as f:
-            json.dump(cfg, f, indent=2)
-
-
-def _list_transformers(pipe: DiffusionPipeline) -> list[tuple[str, torch.nn.Module]]:
-    """Return [(attr_name, module)] for every DiT on the pipeline.
-
-    Wan2.2 A14B has both ``transformer`` (low-noise) and ``transformer_2``
-    (high-noise). TI2V-5B has only ``transformer``.
-    """
-    out = []
-    for attr in ("transformer", "transformer_2"):
-        mod = getattr(pipe, attr, None)
-        if mod is not None:
-            out.append((attr, mod))
-    return out
-
-
-def _save_pipeline_with_fp8_transformer(
-    pipe: DiffusionPipeline,
-    model_path: str,
-    output_dir: Path,
-    max_shard_size: str = "5GB",
-) -> None:
-    """Copy source dir verbatim minus transformer dirs, then save each quantized DiT.
-
-    Handles A14B's dual transformer (``transformer`` + ``transformer_2``) as well
-    as the single-DiT TI2V-5B.
-    """
-    from modelopt.torch.export.diffusers_utils import hide_quantizers_from_state_dict
-
-    src = Path(model_path)
-    if not src.exists():
-        from huggingface_hub import snapshot_download
-
-        src = Path(snapshot_download(model_path))
-
-    if output_dir.exists():
-        shutil.rmtree(output_dir)
-    shutil.copytree(src, output_dir, ignore=shutil.ignore_patterns("transformer", "transformer_2"))
-
-    for attr, backbone in _list_transformers(pipe):
-        transformer_out = output_dir / attr
-        # Pass the nn.Module (transformer), not the Pipeline wrapper.
-        with hide_quantizers_from_state_dict(backbone):
-            backbone.save_pretrained(
-                str(transformer_out),
-                safe_serialization=True,
-                max_shard_size=max_shard_size,
-            )
-
-
-def main() -> None:
-    args = _build_parser().parse_args()
-    if not torch.cuda.is_available():
-        raise SystemExit("CUDA is required for ModelOpt FP8 quantization.")
-
-    mtq = _require_modelopt()
-    model_path, output_dir = _ensure_paths(args)
-    dtype = _select_dtype(args.dtype)
-    prompts = _build_prompts(args)
-    weight_block_size = _parse_block_size(args.weight_block_size)
-
-    print("Quantization plan:")
-    print(f"  input:           {args.model}")
-    print(f"  output:          {output_dir}")
-    print(f"  dtype:           {dtype}")
-    print(f"  height/width:    {args.height}x{args.width}")
-    print(f"  num_frames:      {args.num_frames}")
-    print(f"  calib_size:      {len(prompts)}")
-    print(f"  calib_steps:     {args.calib_steps}")
-    print(f"  quantize_mha:    {args.quantize_mha}")
-    print(
-        f"  weight strategy: {'block-wise ' + str(weight_block_size) if weight_block_size else 'per-tensor (default)'}"
-    )
-
-    pipe = _load_pipeline(model_path, dtype)
-    transformers = _list_transformers(pipe)
-    if not transformers:
-        raise SystemExit("Pipeline has no transformer or transformer_2 attribute.")
-    print(f"  found {len(transformers)} transformer(s): {', '.join(a for a, _ in transformers)}")
-
-    quant_config = copy.deepcopy(mtq.FP8_DEFAULT_CFG)
-    if weight_block_size is not None:
-        quant_config["quant_cfg"]["*weight_quantizer"] = {
-            "num_bits": (4, 3),
-            "block_sizes": {-1: weight_block_size[1], -2: weight_block_size[0]},
-        }
-        print(
-            f"  -> overriding weight quantizer with block_sizes={weight_block_size} "
-            f"({weight_block_size[0]}x{weight_block_size[1]} tiles)"
-        )
-
-    forward_loop = _build_forward_loop(pipe, args, prompts)
-
-    # Quantize each DiT. The forward_loop runs the full pipeline, so quantizers
-    # on every attached transformer (A14B: transformer + transformer_2) get
-    # calibrated during the same passes.
-    for attr, backbone in transformers:
-        print(f"\nQuantizing {attr}...")
-        quantized = mtq.quantize(backbone, quant_config, forward_loop)
-        if quantized is not None:
-            setattr(pipe, attr, quantized)
-        _disable_known_problematic_quantizers(mtq, getattr(pipe, attr), quantize_mha=args.quantize_mha)
-
-    # Export weights to FP8 only AFTER all transformers are calibrated. Exporting
-    # inside the loop would convert transformer's weights to real FP8, and the
-    # next transformer's forward_loop (full pipeline) would then fake-quantize an
-    # already-FP8 weight -> "fake_e4m3fy not implemented for Float8_e4m3fn".
-    print("\nForcing FP8 weight serialization (Wan2.2 isn't in ModelOpt's recognized-model registry,")
-    print("so we have to call the per-weight export helper ourselves)...")
-    total_exported = 0
-    for attr, backbone in _list_transformers(pipe):
-        exported = _force_export_quantized_weights(backbone, dtype)
-        print(f"  -> {exported} weights converted to FP8 in {attr}")
-        total_exported += exported
-
-    if total_exported == 0:
-        raise SystemExit(
-            "No quantized weights were exported. Calibration may have skipped every layer "
-            "(check the disable_quantizer regex) or `mtq.quantize` did not actually wrap any "
-            "weight quantizers."
-        )
-
-    print("\nSaving pipeline with FP8 transformer(s)...")
-    _save_pipeline_with_fp8_transformer(pipe, model_path, output_dir)
-    _patch_quant_config(output_dir, weight_block_size=weight_block_size)
-    print(f"Saved to: {output_dir}")
-    _summarize_export(output_dir)
-
-    print("\nNext: validate the checkpoint with vllm-omni:")
-    print(
-        "  python examples/offline_inference/text_to_video/text_to_video.py \\\n"
-        f"    --model {output_dir} \\\n"
-        "    --quantization fp8 \\\n"
-        "    --prompt 'A dog running across a field of golden wheat.' \\\n"
-        f"    --height {args.height} --width {args.width} --num-frames {args.num_frames} \\\n"
-        "    --num-inference-steps 30 --guidance-scale 5.0 --seed 42 \\\n"
-        "    --output outputs/wan22_modelopt_fp8.mp4"
-    )
-    print(
-        "\n  (--quantization fp8 is auto-upgraded to ModelOpt FP8 at runtime because the "
-        "checkpoint's config.json has modelopt metadata.)"
-    )
-
-
-if __name__ == "__main__":
-    main()
@@ -89,6 +89,14 @@ local = [
    "FlagEmbedding",
 ]

+# LongCat-Video-Avatar A2V/AI2V runtime dependencies for vocal
+# separation and loudness normalization.
+longcat-video-avatar = [
+    "audio-separator==0.44.2",
+    "pyloudnorm==0.1.1",
+    "soundfile>=0.13.1",
+]
+
 # Optional Blackwell-only fused-bias FP8 GEMM; package is `quack-kernels` (imported
 # as `quack`), not the unrelated `quack`. See docs/user_guide/quantization/fp8.md.
 quack = [
@@ -0,0 +1,169 @@
+# MiniMax-H3 on RTX 4090
+
+This recipe uses BF16 weights, tiled VAE decode, tensor parallelism, optional
+Ulysses sequence parallelism, and distributed layerwise offload (DLO). It is the
+24 GiB sibling of [MiniMax-H3-5090.md](MiniMax-H3-5090.md): the two-GPU topology
+matches that recipe, but the resident DiT-block count is lower so the per-rank
+peak fits in 24 GiB. Lower resident counts reduce HBM use and increase
+CPU-to-GPU transfer time.
+
+## Capacity requirements
+
+| Resource | Two RTX 4090s | Four RTX 4090s |
+| --- | ---: | ---: |
+| GPU HBM | 24 GiB per GPU | 24 GiB per GPU |
+| Checkpoint storage | 135 GiB per partition | 135 GiB per partition |
+| Available system RAM | 200 GiB minimum | 200 GiB minimum |
+| Recommended system RAM | 384 GiB | 384 GiB |
+
+`FL2VA` and `Ref2VA` are separate 135 GiB checkpoint partitions. Start one
+server at a time. DLO keeps rank-local weights in pinned host memory; increasing
+`--dlo-resident-layers` improves latency but does **not** reduce host RAM in the
+current implementation because resident layers retain pinned CPU master copies.
+
+A single RTX 4090 is not covered here. The one-GPU DLO profile in the RTX 5090
+recipe peaked at 26.50 GiB with 12 resident layers, which exceeds 24 GiB. On one
+4090, use the model-level CPU offload command in
+[MiniMax-H3.md](MiniMax-H3.md#single-gpu-accuracy-and-memory-first) instead, or
+lower the resident count and re-measure before trusting it.
+
+> **Modular H3:** after #5720 lands, preserve this recipe's one-partition
+> behavior with `--task-type fl2va` or `--task-type ref2va`.
+
+## Two RTX 4090s: 1024x576, 5 seconds
+
+Use TP2 and 12 resident DiT layers. This is the `rtx4090` profile from the main
+recipe. On two RTX 4090s it peaked at 15,620 MiB per rank for T2VA and
+14,918 MiB per rank for Ref2VA, leaving roughly 9 GiB of headroom per card. See
+[Target-hardware validation](#target-hardware-validation) for the full numbers.
+
+```bash
+export VLLM_WORKER_MULTIPROC_METHOD=spawn
+export VLLM_OMNI_VIDEO_SYNC_TIMEOUT=14400
+
+CUDA_VISIBLE_DEVICES=0,1 vllm serve /path/to/MiniMax-H3/FL2VA \
+  --omni --trust-remote-code --host 0.0.0.0 --port 8000 \
+  --num-gpus 2 --tensor-parallel-size 2 --text-encoder-tp-size 2 \
+  --usp 1 --ring 1 --vae-patch-parallel-size 2 \
+  --vae-parallel-mode tile --vae-use-tiling \
+  --enable-distributed-layerwise-offload --dlo-no-use-allgather \
+  --dlo-resident-layers 12 --enforce-eager \
+  --diffusion-attention-backend CUDNN_ATTN
+```
+
+For Ref2VA, stop the FL2VA server and restart the same command with
+`/path/to/MiniMax-H3/Ref2VA`. Ref2VA reference image count and prompt length can
+increase activation memory; begin with one request at a time.
+
+`VLLM_OMNI_VIDEO_SYNC_TIMEOUT` matters more here than on larger cards. DLO
+streams non-resident DiT blocks over PCIe on every denoising step, so a 60-step
+request takes several minutes and would otherwise hit the 600-second default.
+
+## Four RTX 4090s: 1024x576, 5 seconds
+
+Keep TP2 and raise Ulysses sequence parallel to 2 so the world size is
+`TP × USP = 4`. Text-encoder TP and VAE patch parallel follow the GPU count.
+Resident layers stay at 12: adding USP does not shard DiT weights further, so
+per-GPU peak HBM stays close to the two-GPU profile while wall-clock improves.
+
+```bash
+export VLLM_WORKER_MULTIPROC_METHOD=spawn
+export VLLM_OMNI_VIDEO_SYNC_TIMEOUT=14400
+
+CUDA_VISIBLE_DEVICES=0,1,2,3 vllm serve /path/to/MiniMax-H3/FL2VA \
+  --omni --trust-remote-code --host 0.0.0.0 --port 8000 \
+  --num-gpus 4 --tensor-parallel-size 2 --text-encoder-tp-size 4 \
+  --usp 2 --ring 1 --vae-patch-parallel-size 4 \
+  --vae-parallel-mode tile --vae-use-tiling \
+  --enable-distributed-layerwise-offload --dlo-no-use-allgather \
+  --dlo-resident-layers 12 --enforce-eager \
+  --diffusion-attention-backend CUDNN_ATTN
+```
+
+For Ref2VA, stop the FL2VA server and restart the same command with
+`/path/to/MiniMax-H3/Ref2VA`. Pin four dedicated GPUs; if the launcher exposes
+fewer than four devices, startup fails with
+`Stage 0 requires 4 device(s) based on parallel_config`.
+
+## Target-hardware validation
+
+Measured on RTX 4090 (24,564 MiB each, driver 580.126.09) with the serve
+commands above, at vLLM-Omni `0.26.1.dev55+g81b48e83e`, vLLM `0.26.0`, and
+PyTorch `2.11.0+cu130`. GPUs used by each server were dedicated. `Client E2E`
+and `Peak per GPU` come from the `/v1/videos/sync` response headers
+`x-inference-time-s` and `x-peak-memory-mb`; the latter is the rank-0 CUDA
+reserved high-water mark, so it is a per-GPU figure.
+
+| GPUs | Topology | Task | Shape | Frames | Steps | Client E2E | Peak per GPU | Output validation |
+| ---: | --- | --- | ---: | ---: | ---: | ---: | ---: | --- |
+| 2 | TP2 × USP1 | T2VA (`FL2VA`) | 1024x576 | 124 at 24 FPS | 60 | 7 min 9 s | 15.3 GiB | H.264 video + 32 kHz stereo AAC; full `ffmpeg` decode passed |
+| 2 | TP2 × USP1 | Ref2VA | 1024x576 | 124 at 24 FPS | 60 | 14 min 52 s | 14.6 GiB | H.264 video + 32 kHz stereo AAC; full `ffmpeg` decode passed |
+| 4 | TP2 × USP2 | T2VA (`FL2VA`) | 1024x576 | 124 at 24 FPS | 60 | 4 min 29 s | 15.2 GiB | H.264 video + 32 kHz stereo AAC; full `ffmpeg` decode passed |
+| 4 | TP2 × USP2 | Ref2VA | 1024x576 | 124 at 24 FPS | 60 | 9 min 5 s | 16.1 GiB | H.264 video + 32 kHz stereo AAC; successful `/v1/videos/sync` 200 |
+
+All rows use `seed=1101`, `flow_shift=12`, and `audio_flow_shift=3.0`.
+
+Two-GPU repeats: T2VA measured 434.9 s and 429.3 s; Ref2VA measured 891.9 s and
+892.1 s. Peak memory was identical across repeats for both tasks, and the two
+outputs of each pair agreed to within 115 bytes.
+
+Four-GPU T2VA repeats: 273.8 s and 269.5 s, peaking at 15,560 MiB and
+15,612 MiB. Four-GPU Ref2VA is a single successful end-to-end run (545.2 s,
+16,448 MiB). Follow-up Ref2VA requests on this build completed diffusion but
+then hit the engine's hardcoded 30 s async output wait
+(`_ASYNC_OUTPUT_TIMEOUT` in `diffusion_engine.py`) during post-compute D2H,
+so they are not reported as repeats.
+
+Relative to the two-GPU baseline on the same shape, four-GPU TP2×USP2 was about
+1.6× faster for both tasks, while per-GPU peak HBM stayed in the mid-15 GiB
+range because USP does not further shard resident DiT weights. These are
+single-request validation runs, not concurrent throughput benchmarks.
+
+## Request examples
+
+`t2va` requires an explicit `aspect_ratio`. Without it the request fails with
+`t2va requires an explicit aspect_ratio`, even when `width` and `height` are set.
+
+Text to video and audio, against the `FL2VA` server:
+
+```bash
+curl -sS -D headers_t2va.txt -o out_t2va.mp4 --max-time 14400 \
+  -X POST "http://127.0.0.1:8000/v1/videos/sync" \
+  -F 'prompt=傍晚的小厨房，夕阳余晖从窗边洒进来，旧木桌上放着洗到一半的马克杯和起雾的玻璃瓶，悬挂的抹布轻轻晃动。画面带有手持拍摄的轻微晃动和逆光曝光波动，生活感十足。环境音是安静的厨房底噪。' \
+  -F 'aspect_ratio=16:9' \
+  -F 'width=1024' -F 'height=576' -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}'
+```
+
+Reference to video and audio, against the `Ref2VA` server:
+
+```bash
+curl -sS -D headers_ref2va.txt -o out_ref2va.mp4 --max-time 14400 \
+  -X POST "http://127.0.0.1:8000/v1/videos/sync" \
+  -F 'prompt=傍晚的小厨房，夕阳余晖从窗边洒进来，旧木桌上放着洗到一半的马克杯和起雾的玻璃瓶，悬挂的抹布轻轻晃动。画面带有手持拍摄的轻微晃动和逆光曝光波动，生活感十足。环境音是安静的厨房底噪。' \
+  -F 'input_reference=@/path/to/reference.jpg;type=image/jpeg' \
+  -F 'audio_reference=</path/to/audio_reference.json' \
+  -F 'width=1024' -F 'height=576' -F 'fps=24' \
+  -F 'num_inference_steps=60' -F 'flow_shift=12' -F 'seed=1101' \
+  -F 'extra_params={"task":"ref2va","duration":5,"audio_flow_shift":3.0}'
+```
+
+`audio_reference` is a JSON object, which is why the example loads it from a file
+with curl's `<` syntax. Its `audio_url` must be an `http(s)` URL or a data URL;
+a bare filesystem path is rejected with `Invalid audio_reference.audio_url`.
+
+```json
+{"audio_url": "https://example.com/reference.wav"}
+```
+
+Reference images are validated before inference: the short edge must be at least
+256 pixels, the long edge at most 5760, and the aspect ratio must fall between
+0.4 and 2.5.
+
+The response headers carry the measurements used in the table above:
+
+```bash
+cat headers_t2va.txt
+ffmpeg -v error -i out_t2va.mp4 -f null - && echo DECODE_OK
+```
@@ -79,7 +79,8 @@ sharing the tokenizer, processor, text encoder, and VAEs from `FL2VA`.
 ### Multi-NPU: 768P combined-service configuration

 Use Ulysses sequence parallelism degree 8, text-encoder tensor parallelism
-degree 8, native tiled VAE patch parallelism degree 8, and layerwise offload:
+degree 8, native tiled VAE patch parallelism degree 8, and distributed
+layerwise offload:

 ```bash
 export ASCEND_RT_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
@@ -98,7 +99,7 @@ vllm serve "${MODEL}" \
  --usp 8 \
  --ring 1 \
  --text-encoder-tp-size 8 \
-  --enable-layerwise-offload \
+  --enable-distributed-layerwise-offload \
  --vae-parallel-mode tile \
  --vae-use-tiling \
  --vae-patch-parallel-size 8 \
@@ -112,6 +113,11 @@ H3 is CFG-distilled, so `--cfg-parallel-size` must remain 1.
 The same endpoint accepts `task=t2va`, `task=fl2va`, and `task=ref2va`; no
 partition restart is required. Layerwise offload applies to both DiTs.

+On Atlas 800I A3 (64 GB HBM per device) the combined service does not fit at
+768P without offloading or sharding: use distributed layerwise offload (as
+above) or HSDP — see
+[§ Memory and attention optimizations](#memory-and-attention-optimizations-a3).
+
 ### Optional optimizations

 Two independent optimizations may be enabled on top of the configuration
@@ -134,6 +140,63 @@ Keep `--ring 1` when using RainFusion: the `rf_v2` kernel ranks key blocks
 over the whole sequence, so ring parallelism would split away the keys it
 needs. Scale with `--usp` instead.

+## Memory and attention optimizations (A3)
+
+### Fitting 768P into 64 GB HBM: distributed layerwise offload or HSDP
+
+Each Atlas 800I A3 NPU has 64 GB of HBM, which is not enough for the
+combined MiniMax-H3 service at 768P. Enable **one** of the following at
+server startup:
+
+**Distributed layerwise offload** — DiT layers are offloaded with parameters
+gathered across the parallel group instead of replicated per rank:
+
+```bash
+  --enable-distributed-layerwise-offload
+```
+
+Measured peak NPU memory is about 45 GB per device for Ref2VA with a 13.88 s
+input video generating a 15 s 768P (1344x768) output video. Host (CPU) memory
+usage is high with this option.
+
+**HSDP** — hybrid sharded data parallelism for the DiT parameters. The
+multi-stream memory-reuse knob is mandatory with this flag:
+
+```bash
+export MULTI_STREAM_MEMORY_REUSE=2
+```
+
+```bash
+  --use-hsdp
+```
+
+Host memory usage is small, but large shapes may still OOM; HBM usage
+optimizations for this path are ongoing.
+
+### FLASH_ATTN backend with MindIE-SD
+
+Keep `--diffusion-attention-backend FLASH_ATTN` and install MindIE-SD (see
+[§ Environment](#environment)). On NPU this backend carries most of the
+memory-reduction and performance work: a mask-free packed varlen path that
+never materializes the quadratic `full_qk` padding mask, and K/V prefix
+slicing, both driven by the packed `cu_seqlens` metadata emitted by the H3
+transformer.
+
+### Optional: LaserAttention fused kernel
+
+For an additional attention speedup, select the Ascend LaserAttention fused
+kernel before starting the server:
+
+```bash
+export MINDIE_SD_FA_TYPE="ascend_laser_attention"
+```
+
+This requires the FLASH_ATTN backend and MindIE-SD. H3 automatically applies
+exact power-of-two input pre-scaling (`laser_input_scale=256`) so the
+kernel's fp16 workspace cannot overflow on outlier activations. Measured
+speedup numbers will be added here.
+
+
 ## HTTP API examples

 The request format is identical to the GPU recipe; see
@@ -154,8 +217,8 @@ configuration above:

 | Workload | Configuration |
 |----------|---------------|
-| T2VA, 209 frames, 1344x768 | TE TP8, layerwise offload, Ulysses 8, VPP8 tile, regional compile |
-| Ref2VA (prompt + video), 124 frames, 1344x768 | TE TP8, layerwise offload, Ulysses 8, VPP8 tile, regional compile |
+| T2VA, 209 frames, 1344x768 | TE TP8, distributed layerwise offload, Ulysses 8, VPP8 tile, regional compile |
+| Ref2VA (prompt + video), 124 frames, 1344x768 | TE TP8, distributed layerwise offload, Ulysses 8, VPP8 tile, regional compile |

 These measurements describe the validated shapes rather than a general
 throughput guarantee.
@@ -0,0 +1,346 @@
+# MiniMax-H3 on RTX PRO 6000 Blackwell GPUs
+
+This recipe runs MiniMax-H3 in BF16 on 96 GiB RTX PRO 6000 Blackwell GPUs.
+The additional HBM over the RTX PRO 5000 profile removes the four-GPU
+minimum: two GPUs are enough to keep the model resident with TP2. Four and
+eight GPUs raise the Ulysses degree to shard the attention sequence further.
+CPU offload and distributed layerwise offload are not required in any of
+these configurations.
+
+Validated on:
+- Host: YLX Y762
+- GPUs: 8 × RTX PRO 6000 Blackwell (96 GiB)
+- Device order: default (`CUDA_VISIBLE_DEVICES` not set)
+- Driver / CUDA: 580.105.08 / 13.0
+- Container image: `vllm/vllm-omni:minimax-h3`
+- Workload: T2VA, 1344×768, `duration=5.0`, 50 steps, `flow_shift=12`, `seed=1101`
+
+## Capacity requirements
+
+| Resource | Two GPUs | Four GPUs | Eight GPUs |
+| --- | ---: | ---: | ---: |
+| GPU HBM | 96 GiB per GPU | 96 GiB per GPU | 96 GiB per GPU |
+| Checkpoint storage | 135 GiB per partition | 135 GiB per partition | 135 GiB per partition |
+| Available system RAM | 200 GiB minimum | 200 GiB minimum | 200 GiB minimum |
+| Recommended system RAM | 384 GiB | 384 GiB | 384 GiB |
+| Measured peak HBM per GPU | 77.49 GiB | 66.44 GiB | 61.07 GiB |
+
+`FL2VA` and `Ref2VA` are separate checkpoint partitions. Start one server at
+a time on a host sized for the minimum system-memory requirement, or pass
+`--task-type fl2va` or `--task-type ref2va` to download and load a single
+partition. The two-server layout described in the eight-GPU section doubles
+both the storage and the host-memory requirement.
+
+## Parallelism summary
+
+| Configuration | TP | Ulysses | Text-encoder TP | VAE patch parallel |
+| --- | ---: | ---: | ---: | ---: |
+| Two GPUs | 2 | 1 | 2 | 2 |
+| Four GPUs | 2 | 2 | 4 | 4 |
+| Eight GPUs | 2 | 4 | 8 | 8 |
+| Eight GPUs, headroom variant | 4 | 2 | 8 | 8 |
+
+Per-GPU weight residency is set by the tensor-parallel degree alone. Raising
+the Ulysses degree lowers activation memory and per-step latency but leaves
+weight residency unchanged. Ring parallelism stays at 1, and H3 is
+CFG-distilled, so `--cfg-parallel-size` must also remain 1.
+
+## PCIe topology and GPU order
+
+RTX PRO 6000 Blackwell does not provide NVLink. Before starting the server,
+inspect the host topology:
+
+```bash
+nvidia-smi topo -m
+nvidia-smi nvlink -s
+```
+
+Tensor-parallel groups are consecutive ranks. Ulysses groups are strided by
+the tensor-parallel degree. Confirm this grouping against the server log
+before tuning the device order on a new host.
+
+For the two-GPU configuration, pick a single `PXB` pair on one NUMA node.
+For the four-GPU configuration, pick two `PXB` pairs on the same node and
+order them so that each Ulysses group lands inside one pair. If physical
+GPUs `(0,1)` and `(2,3)` are the two local pairs, the order is
+`CUDA_VISIBLE_DEVICES=0,2,1,3`. Do not copy these IDs blindly: reproduce the
+same relationship on the target host.
+
+Device-order pinning and NUMA binding are tuning knobs, not prerequisites.
+The measurements in this recipe were taken without either, so any gain from
+them is additional to the numbers reported below and is host-specific.
+
+## Two-GPU serving configuration
+
+Two 96 GiB GPUs hold the BF16 model with TP2 alone. There is no Ulysses
+group, so the DiT attention sequence is not sharded and activation memory is
+higher than in the four-GPU profile at the same output shape. Re-measure
+peak memory before increasing the output shape, the reference-input length,
+or concurrency.
+
+```bash
+export MODEL_ROOT=/path/to/MiniMax-H3
+export MODEL="${MODEL_ROOT}/FL2VA"
+export PORT=8000
+
+VLLM_OMNI_VIDEO_SYNC_TIMEOUT=1800 \
+vllm serve "${MODEL}" \
+  --omni \
+  --host 0.0.0.0 \
+  --port "${PORT}" \
+  --trust-remote-code \
+  --num-gpus 2 \
+  --tensor-parallel-size 2 \
+  --usp 1 \
+  --ring 1 \
+  --text-encoder-tp-size 2 \
+  --vae-patch-parallel-size 2 \
+  --vae-parallel-mode tile \
+  --vae-use-tiling \
+  --diffusion-attention-backend CUDNN_ATTN
+```
+
+## Four-GPU serving configuration
+
+Four GPUs keep TP2 for the weight shard and add Ulysses2, which shards the
+attention sequence and lowers per-GPU activation memory. Text-encoder TP4
+and VAE patch parallelism 4 spread the Qwen3-VL encoder and tiled decode
+across all four GPUs. TP1 is not an option on this card: an unsharded BF16
+partition does not fit in 96 GiB.
+
+```bash
+export MODEL_ROOT=/path/to/MiniMax-H3
+export MODEL="${MODEL_ROOT}/FL2VA"
+export PORT=8000
+
+VLLM_OMNI_VIDEO_SYNC_TIMEOUT=1800 \
+vllm serve "${MODEL}" \
+  --omni \
+  --host 0.0.0.0 \
+  --port "${PORT}" \
+  --trust-remote-code \
+  --num-gpus 4 \
+  --tensor-parallel-size 2 \
+  --usp 2 \
+  --ring 1 \
+  --text-encoder-tp-size 4 \
+  --vae-patch-parallel-size 4 \
+  --vae-parallel-mode tile \
+  --vae-use-tiling \
+  --diffusion-attention-backend CUDNN_ATTN
+```
+
+## Eight-GPU serving configuration
+
+Eight GPUs extend the Ulysses degree to 4 while keeping TP2. This does not
+lower per-GPU weight residency relative to the four-GPU profile; it lowers
+activation memory and per-step latency by sharding the attention sequence
+four ways.
+
+```bash
+export MODEL_ROOT=/path/to/MiniMax-H3
+export MODEL="${MODEL_ROOT}/FL2VA"
+export PORT=8000
+
+VLLM_OMNI_VIDEO_SYNC_TIMEOUT=1800 \
+vllm serve "${MODEL}" \
+  --omni \
+  --host 0.0.0.0 \
+  --port "${PORT}" \
+  --trust-remote-code \
+  --num-gpus 8 \
+  --tensor-parallel-size 2 \
+  --usp 4 \
+  --ring 1 \
+  --text-encoder-tp-size 8 \
+  --vae-patch-parallel-size 8 \
+  --vae-parallel-mode tile \
+  --vae-use-tiling \
+  --diffusion-attention-backend CUDNN_ATTN
+```
+
+### Rank ordering across two sockets
+
+With TP2 and Ulysses4 the two Ulysses groups are ranks `(0,2,4,6)` and
+`(1,3,5,7)`, and the four tensor-parallel pairs are `(0,1)`, `(2,3)`,
+`(4,5)`, and `(6,7)`. On a dual-socket host these two collectives cannot
+both be socket-local. If you choose to pin the device order, the trade-off
+is:
+
+| Device order | Socket-local collective | Crosses sockets |
+| --- | --- | --- |
+| `0,4,1,5,2,6,3,7` | Ulysses all-to-all | TP all-reduce |
+| `0,1,2,3,4,5,6,7` | TP all-reduce | Ulysses all-to-all |
+
+The first order assumes physical GPUs `0-3` are on NUMA node 0 and `4-7` on
+node 1. Reproduce that relationship on the target host rather than copying
+the IDs. The reported measurements use the default order and no NUMA
+binding; the two orders were not compared on this host.
+
+### Headroom variant: TP4 with Ulysses2
+
+Raising the tensor-parallel degree to 4 halves per-GPU weight residency and
+leaves room for long Ref2VA references, larger output shapes, or
+concurrency greater than one. It also adds a four-way all-reduce per DiT
+block over a link without NVLink. Use it when the TP2 profile runs out of
+headroom, not as the default. Replace the parallel flags above with:
+
+```bash
+  --tensor-parallel-size 4 \
+  --usp 2 \
+```
+
+### Two independent four-GPU servers
+
+An eight-GPU host can instead run one FL2VA server and one Ref2VA server
+side by side, each pinned to its own NUMA node with the four-GPU
+configuration. This removes every cross-socket collective and lifts the
+one-server-at-a-time restriction, at the cost of doubling the requirements
+to 400 GiB available system RAM and 270 GiB of model storage.
+
+Use `CUDA_VISIBLE_DEVICES=0,2,1,3`, `--cpunodebind=0 --membind=0`, and
+`--port 8091` for the FL2VA server. Use `CUDA_VISIBLE_DEVICES=4,6,5,7`,
+`--cpunodebind=1 --membind=1`, and `--port 8092` for the Ref2VA server.
+This layout is described for completeness and was not part of the measured
+runs.
+
+## Shared serving notes
+
+Do not add `--enforce-eager` for a performance run. Warm the server before
+measuring so regional compilation is outside the measured request. On this
+host the first request after startup ran 19% slower than the steady state
+(107.58 s versus 90.48 s on eight GPUs); two warmup requests were enough to
+converge.
+
+For Ref2VA on a single-server layout, stop the FL2VA server and restart the
+same command with `MODEL="${MODEL_ROOT}/Ref2VA"`.
+
+## Attention backend
+
+RTX PRO 6000 Blackwell is an SM120 part, so the datacenter-Blackwell
+`TRTLLM_ATTN` default does not apply here. Every configuration above selects
+cuDNN BF16 attention explicitly. This keeps the recipe independent of
+platform-default backend changes and does not depend on the experimental
+SM120 kernel.
+
+## Optional: online FP8
+
+`--quantization fp8` quantizes eligible DiT linears at load time and is
+compatible with tensor parallelism and VAE tiling. Use it when the resident
+BF16 peak leaves too little headroom for long Ref2VA references, a larger
+output shape, or concurrency greater than one. It cannot be combined with
+layerwise offload.
+
+## Target-hardware validation
+
+T2VA, 1344×768, `duration=5.0`, `fps=24`, 50 steps, `flow_shift=12`,
+`seed=1101`, BF16, `CUDNN_ATTN`, tiled VAE, one request at a time. Servers
+were started with the commands above — default device order, no NUMA
+binding — plus `--enable-diffusion-pipeline-profiler`. Two warmup requests
+preceded each measured request.
+
+| Measurement | Two GPUs | Four GPUs | Eight GPUs |
+| --- | ---: | ---: | ---: |
+| Topology | TP2 × USP1 | TP2 × USP2 | TP2 × USP4 |
+| Text encode | 0.040 s | 0.044 s | 0.035 s |
+| Denoise, 50 steps | 278.55 s | 168.73 s | 87.90 s |
+| Per step | 5.571 s | 3.375 s | 1.758 s |
+| VAE decode | 5.396 s | 2.791 s | 1.798 s |
+| Client E2E | 284.76 s | 172.32 s | 90.48 s |
+| Peak HBM per GPU | 77.49 GiB | 66.44 GiB | 61.07 GiB |
+| Headroom below 96 GiB | 18.5 GiB | 29.6 GiB | 34.9 GiB |
+
+Stage times are read from the `X-Stage-Durations` response header of
+`/v1/videos/sync`. Peak memory is the maximum of `nvidia-smi
+--query-gpu=memory.used` sampled at 1 Hz across every device for the
+duration of the measured request. Per step is denoise wall time divided by
+the 50 requested steps. Stage times sum to roughly 0.75 s less than
+end-to-end in all three configurations; queueing, result transfer, and MP4
+muxing sit outside the profiled stages.
+
+### Scaling behaviour
+
+Denoise speedup is not uniform across the sweep: 1.65× from two to four
+GPUs (83% efficiency) but 1.92× from four to eight (96%). Two GPUs run
+Ulysses1, so there is no sequence parallelism at all; going to four GPUs
+pays the one-time cost of introducing the all-to-all, while going from four
+to eight only widens an all-to-all that already exists. Expect the second
+doubling to pay off better than the first on this topology.
+
+VAE decode behaves in the opposite direction — 1.93× from two to four GPUs
+but only 1.55× from four to eight — as patch parallelism produces smaller
+tiles and tile-boundary overhead grows.
+
+Text encoding is 0.04 s in every configuration and is not worth optimising;
+denoise is 96-98% of end-to-end throughout.
+
+### Per-GPU memory model
+
+Peak memory falls by 11.05 GiB from two to four GPUs and by a further
+5.37 GiB from four to eight — halving each time the Ulysses degree doubles,
+because all three configurations shard the DiT with TP2 and hold identical
+weights per GPU. Fitting the first two points gives:
+
+```
+peak HBM per GPU ≈ 55.4 GiB + 22.1 GiB / ulysses_degree
+```
+
+That fit predicts 60.9 GiB at Ulysses4 against 61.07 GiB measured, a 0.25%
+error. The practical consequence is a floor near 55 GiB per GPU: adding
+GPUs past eight will not reduce the per-GPU requirement further. To serve
+on smaller cards, raise the tensor-parallel degree instead.
+
+Re-measure memory for longer reference inputs, concurrency greater than
+one, or a different output shape.
+
+### Not yet measured
+
+Ref2VA latency and memory have not been measured on this host. Ref2VA
+drives the Qwen3-VL presentation from tens of tokens for text-only prompts
+to several thousand with image or video references, so T2VA numbers do not
+transfer to it. The eight-GPU device-order comparison in the section above
+was also not run.
+
+## T2VA request example
+
+`t2va` requires an explicit `aspect_ratio` even when `width` and `height`
+are supplied.
+
+```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=1344' \
+  -F 'height=768' \
+  -F 'aspect_ratio=16:9' \
+  -F 'fps=24' \
+  -F 'num_inference_steps=50' \
+  -F 'flow_shift=12' \
+  -F 'seed=1101' \
+  -F 'extra_params={"task":"t2va","duration":5.0,"audio_flow_shift":3.0}' \
+  -o t2va.mp4
+```
+
+To collect the stage breakdown and peak memory for a request, add
+`--enable-diffusion-pipeline-profiler` to the server command and `-D
+headers.txt` to the curl invocation, then read `X-Stage-Durations` and
+`X-Peak-Memory-MB` from the saved headers.
+
+## Ref2VA request example
+
+`ref2va` defaults to a 16:9 aspect ratio, so `aspect_ratio` is optional
+here.
+
+```bash
+curl -X POST "http://127.0.0.1:${PORT}/v1/videos/sync" \
+  --fail-with-body -w '\nHTTP %{http_code}\n' \
+  --max-time 1200 \
+  -F "input_reference=@/root/hand.jpg;type=image/jpeg" \
+  -F "audio_reference=</root/audio_ref.json" \
+  -F "prompt=2D动画融合在一起的影像。夕阳余晖残留在窗边，生活感十足的小厨房里有旧木桌、洗到一半的马克杯、起雾的玻璃瓶、悬挂的抹布。画面带有智能手机单手拍摄的手抖、近距离对焦的犹豫、逆光曝光波动。要像在家中慌忙拍下某个不可思议事件的自然质感，不要广告影像的精心整理。声音只用厨房环境声与手绘生物柔和的电子音、小小的叫声。" \
+  -F 'width=1344' -F 'height=768' -F 'fps=24' \
+  -F 'num_inference_steps=60' -F 'flow_shift=12' -F 'seed=1101' \
+  -F 'extra_params={"task":"ref2va","duration":8,"audio_flow_shift":3.0}' \
+  -o /root/out_ref2va.mp4
+```
@@ -0,0 +1,262 @@
+# MiniMax-H3 on DGX Spark (GB10)
+
+This recipe uses online FP8 weight quantization, tiled VAE decode, and a single
+resident partition. GB10 is a unified-memory platform, so unlike the discrete-GPU
+recipes it uses **no offload of any kind** — see the capacity note below.
+
+Validated on:
+
+- Host: DGX Spark (GB10), aarch64
+- GPUs: 1 (unified memory)
+- Driver: <FILL_DRIVER_VERSION>
+- vLLM: 0.26.0
+- vLLM-Omni: `main` at `e1aa6eae75c460cd1893bc320546e81e66973831`
+- Workloads: T2VA and Ref2VA, 960x576, `duration=8.0`, 50 steps, `flow_shift=12`,
+  `seed=1101`
+
+
+## Capacity requirements
+
+| Resource | DGX Spark (GB10) |
+| --- | ---: |
+| Unified memory | 128 GB LPDDR5X (~121 GiB usable) |
+| GPUs | 1 |
+| Checkpoint storage | 135 GiB per partition |
+| Measured peak allocator high-water, T2VA (FP8) | 97.7 GiB |
+| Measured peak allocator high-water, Ref2VA (FP8) | 102.8 GiB |
+
+`FL2VA` and `Ref2VA` are separate 135 GiB checkpoint partitions. Start one server
+at a time.
+
+On GB10 the CPU and GPU share one physical memory pool, which changes the capacity
+math relative to the RTX and datacenter recipes:
+
+- **Do not use `--enable-distributed-layerwise-offload`.** DLO stages rank-local
+  weights into pinned host memory, which on GB10 is the same pool the weights are
+  already in. Peak usage roughly doubles and the process is killed by the OOM
+  killer (`Exit code: -9`) shortly after `Enabling offloader backend`.
+- **Do not use `--enable-cpu-offload`.** Moving weights from "VRAM" to "host RAM"
+  frees nothing here.
+- **`--quantization fp8` is mandatory.** A BF16 partition is 135 GiB and does not
+  fit in 121 GiB. Online FP8 quantizes the DiT only (62 GiB to ~31 GiB); the
+  Qwen3-VL text encoder and both VAEs stay BF16.
+
+With ~97.7 GiB of the pool held by the allocator at peak, T2VA leaves roughly
+23 GiB of headroom for the OS, page cache, and any other process on the box.
+Ref2VA is tighter on both counts. The allocator high-water rises to 102.8 GiB,
+and the allocator figure understates true pool usage.
+
+Sampling `/proc/meminfo` at 1 Hz across a 50-step Ref2VA request on a host with
+`MemTotal: 127600524 kB` (121.7 GiB), `MemAvailable` started at 8.1 GiB, fell to
+a trough of **7.4 GiB**, and recovered to 8.9 GiB. At the trough the pool held
+roughly **114.3 GiB** — 11.4 GiB more than the 102.8 GiB the allocator reports,
+which is the CUDA context, non-PyTorch allocations, and page cache the header
+does not count.
+
+Treat this platform as strictly single-tenant for Ref2VA. There is no room for a
+second process, and adding reference images, reference videos, or a larger output
+shape will run into the OOM killer.
+
+## One GB10: 960x576, 8 seconds
+
+```bash
+CUDA_VISIBLE_DEVICES=0 \
+FLASHINFER_DISABLE_VERSION_CHECK=1 \
+VLLM_WORKER_MULTIPROC_METHOD=spawn \
+VLLM_OMNI_VIDEO_SYNC_TIMEOUT=7200 \
+vllm serve /path/to/MiniMax-H3/FL2VA \
+  --omni --trust-remote-code --host 0.0.0.0 --port 8000 \
+  --init-timeout 3600 \
+  --num-gpus 1 --tensor-parallel-size 1 --text-encoder-tp-size 1 \
+  --usp 1 --ring 1 --vae-patch-parallel-size 1 \
+  --vae-parallel-mode tile --vae-use-tiling \
+  --quantization fp8 --enforce-eager \
+  --diffusion-attention-backend CUDNN_ATTN
+```
+
+Wait for `Application startup complete`, then run a short smoke test before
+committing to a full 50-step request.
+
+`t2va` requires an explicit `aspect_ratio` even when `width` and `height` are
+supplied; omitting it fails the request with
+`OmniClientError: t2va requires an explicit aspect_ratio`.
+
+```bash
+curl --fail-with-body -sS -X POST http://127.0.0.1:8000/v1/videos/sync \
+  -F 'prompt=At night, three cats march into a bedroom playing tiny brass instruments, then abruptly file out, with synchronized room ambience.' \
+  -F 'width=960' \
+  -F 'height=576' \
+  -F 'aspect_ratio=16:9' \
+  -F 'fps=24' \
+  -F 'num_inference_steps=10' \
+  -F 'flow_shift=12' \
+  -F 'seed=1101' \
+  -F 'extra_params={"task":"t2va","duration":4.0,"audio_flow_shift":3.0}' \
+  -o h3-gb10-smoke.mp4
+```
+
+The full-quality request uses 50 steps:
+
+```bash
+curl --fail-with-body -sS -X POST http://127.0.0.1:8000/v1/videos/sync \
+  -F 'prompt=At night, three cats march into a bedroom playing tiny brass instruments, then abruptly file out, with synchronized room ambience.' \
+  -F 'width=960' \
+  -F 'height=576' \
+  -F 'aspect_ratio=16:9' \
+  -F 'fps=24' \
+  -F 'num_inference_steps=50' \
+  -F 'flow_shift=12' \
+  -F 'seed=1101' \
+  -F 'extra_params={"task":"t2va","duration":8.0,"audio_flow_shift":3.0}' \
+  -o h3-gb10-t2va.mp4
+
+ffprobe -v error -show_entries \
+  stream=index,codec_name,width,height,r_frame_rate,sample_rate,channels \
+  -of json h3-gb10-t2va.mp4
+```
+
+Both requests return `200 OK` with an MP4 body: H.264 video at the requested
+geometry plus a 32 kHz stereo AAC track.
+
+## One GB10: Ref2VA, 960x576, 8 seconds
+
+`FL2VA` and `Ref2VA` are separate 135 GiB partitions and only one fits at a time.
+Stop the FL2VA server before starting this one.
+
+```bash
+CUDA_VISIBLE_DEVICES=0 \
+FLASHINFER_DISABLE_VERSION_CHECK=1 \
+VLLM_WORKER_MULTIPROC_METHOD=spawn \
+VLLM_OMNI_VIDEO_SYNC_TIMEOUT=14400 \
+vllm serve /path/to/MiniMax-H3/Ref2VA \
+  --omni --trust-remote-code --host 0.0.0.0 --port 8000 \
+  --init-timeout 3600 \
+  --num-gpus 1 --tensor-parallel-size 1 --text-encoder-tp-size 1 \
+  --usp 1 --ring 1 --vae-patch-parallel-size 1 \
+  --vae-parallel-mode tile --vae-use-tiling \
+  --quantization fp8 --enforce-eager \
+  --diffusion-attention-backend CUDNN_ATTN
+```
+
+The 7200 s timeout used for T2VA is too short here: a 50-step Ref2VA request
+takes over an hour on this platform. Use 14400 s.
+
+Unlike `t2va`, `ref2va` defaults to a 16:9 output ratio, so `aspect_ratio` can be
+omitted when `width` and `height` are supplied.
+
+```bash
+curl --fail-with-body -sS -X POST http://127.0.0.1:8000/v1/videos/sync \
+  -D ref2va-headers.txt \
+  -w 'client_e2e_s=%{time_total}\n' \
+  --max-time 14400 \
+  -F 'input_reference=@/path/to/reference.jpg;type=image/jpeg' \
+  -F 'prompt=A hand rests on a wooden table in warm evening light, then slowly opens and turns toward the camera, with quiet room ambience.' \
+  -F 'width=960' \
+  -F 'height=576' \
+  -F 'fps=24' \
+  -F 'num_inference_steps=50' \
+  -F 'flow_shift=12' \
+  -F 'seed=1101' \
+  -F 'extra_params={"task":"ref2va","duration":8.0,"audio_flow_shift":3.0}' \
+  -o h3-gb10-ref2va.mp4
+```
+
+The response is `200 OK` with an MP4 body: H.264 at the requested geometry plus a
+32 kHz stereo AAC track. `ffprobe` reports 8.032 s for `duration=8.0`.
+
+## Measured latency
+
+Single GB10, FP8, eager, cuDNN attention, tiled VAE, one request at a time:
+
+| Workload | Steps | Text encode | Denoise | Per step | VAE decode | MP4 mux | Client E2E | Peak memory |
+| --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
+| T2VA, 4 s, smoke shape | 10 | — | — | 8.68 s | — | 0.49 s | 95.9 s | — |
+| T2VA, 8 s, 960x576 | 50 | 0.25 s | 2088 s | 42.61 s | 70 s | 4.84 s | 2169.4 s (36 min 9 s) | 97.7 GiB |
+| Ref2VA, 8 s, 960x576, one image (3 runs) | 10 | 8.43-10.94 s | 665.5-748.8 s | 66.55-74.88 s | 67.9-69.8 s | 3.63 s | 770.3-861.1 s | 102.8 GiB |
+| Ref2VA, 8 s, 960x576, one image | 50 | 9.66 s | 4039.6 s | 80.79 s | 69.4 s | 3.63 s | 4157.0 s (69 min 17 s) | 102.8 GiB |
+
+How these were measured:
+
+- Stage times come from `--enable-diffusion-pipeline-profiler`, read from the
+  `X-Stage-Durations` response header of `/v1/videos/sync`. The keys are
+  `MiniMaxH3Pipeline.encode_prompt` (text encode),
+  `MiniMaxH3Pipeline.diffuse` (denoise), and `MiniMaxH3Pipeline.decode`
+  (VAE decode).
+- MP4 mux is not a header field. It is the
+  `Video response encoding (MP4 bytes)` line in the server log.
+- Peak memory is the `X-Peak-Memory-MB` header, which reports the CUDA allocator
+  high-water mark (`torch.cuda.max_memory_reserved`) for the request. It excludes
+  the CUDA context and non-PyTorch allocations, so true pool usage is somewhat
+  higher. `nvidia-smi --query-gpu=memory.used` is not usable on GB10 — unified
+  memory has no discrete VRAM to report and the field returns `[N/A]`. See the
+  capacity section for a `/proc/meminfo` cross-check.
+- For T2VA, text encode and VAE decode were measured on a separate 10-step run at
+  the same 960x576 / 8 s geometry. Both are independent of step count, so they
+  transfer to the 50-step row unchanged. The Ref2VA MP4 mux figure transfers the
+  same way, from the 10-step run at identical output geometry and duration.
+- Ref2VA peak memory is byte-identical between the 10-step and 50-step runs
+  (`X-Peak-Memory-MB: 105318.000` in both), confirming that peak is set by output
+  geometry and reference input rather than step count.
+- Stage times do not sum exactly to end-to-end. Queueing, result transfer, and
+  engine bookkeeping sit outside the profiled stages.
+
+Denoise dominates at 96% of end-to-end, so step count and geometry are the only
+knobs that materially change the wall clock on this platform.
+
+Ref2VA costs 1.92x the wall clock of T2VA at the same geometry and step count:
+4157.0 s versus 2169.4 s. Text encoding shows the most dramatic ratio — 9.66 s
+versus 0.25 s, 39x — but at 0.2% of end-to-end it is not what makes Ref2VA slow.
+The cost is in denoise. A single reference image expands the Qwen3-VL
+presentation to 6883 tokens (logged as `MiniMax H3 ref2va Qwen presentation:
+6883 tokens, 1 reference images`), and those tokens sit in the attention context
+of every denoise step. Per-step time rises from 42.61 s to 80.79 s, and that
+1.90x multiplies across all 50 steps. VAE decode is unchanged at ~69 s, as
+expected for an identical output shape.
+
+Repeated identical 10-step T2VA requests varied by up to 23% in denoise time
+(397 s to 490 s) on a machine that had already been under sustained load. Treat
+these as single-run measurements on one machine, not a throughput guarantee, and
+expect the slower end of that range for back-to-back jobs.
+
+Ref2VA per-step time is load-dependent in the same way. On an idle machine the
+first 10-step request ran at 66.55 s per step. The 50-step request that followed
+occupied the box for 69 consecutive minutes and averaged 80.79 s per step. Two
+further 10-step requests taken immediately afterwards settled at 74.88 s and
+74.31 s — 12.5% slower than the idle run, and within 0.8% of each other, so this
+is a stable thermal state rather than measurement noise.
+
+Plan for 81 s per step when sizing a full 50-step Ref2VA job; 66 s per step is
+only reachable on a cold machine and does not hold for the length of one request.
+
+Ref2VA text encode converged downward across the four runs (10.94 s, 9.66 s,
+8.52 s, 8.43 s) as first-touch costs fell away; treat ~8.5 s as steady state.
+VAE decode was stable at 67.9-69.8 s throughout.
+
+Set `VLLM_OMNI_VIDEO_SYNC_TIMEOUT` well above the expected request time — the
+default 1800 s is shorter than a 50-step run on this platform, and 7200 s is
+shorter than a 50-step Ref2VA run.
+
+## Known limitations
+
+- One GPU means no Ulysses, ring, TP, or VAE patch parallelism. Pass
+  `--usp 1 --ring 1 --vae-patch-parallel-size 1` explicitly if a profile or
+  script would otherwise supply higher degrees.
+- `--enforce-eager` is used because regional compile adds startup time without a
+  parallelism win on a single rank.
+- `t2va` rejects requests without an explicit `aspect_ratio`. When `width` and
+  `height` are both supplied they take precedence and `aspect_ratio` only has to
+  name one of the supported ratios.
+- Sustained load degrades throughput. See the variance notes above.
+- Only one partition is resident. For Ref2VA, stop the FL2VA server and restart
+  with `/path/to/MiniMax-H3/Ref2VA`. Ref2VA with one reference image is measured
+  above; reference videos and additional images push the token count higher and
+  were not measured. With `MemAvailable` bottoming out at 7.4 GiB during a
+  single-image request, there is very little room to grow — re-measure before
+  adding references or raising the output shape.
+- Ref2VA requires a vLLM-Omni build newer than `v0.26.0`. See the version note
+  under **Validated on**: FP8 weight loading and image-only Ref2VA are both
+  broken on `release/v0.26.0`.suggestion is  `v0.26.1`.
+- Online FP8 is incompatible with layerwise offload — the offload path produces a
+  weight stride the Cutlass FP8 kernel rejects. This is not a practical
+  restriction here since offload is unusable on GB10 anyway.
+```
@@ -750,8 +750,73 @@ 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.

+## TeaCache acceleration
+
+TeaCache reuses DiT block residuals across denoising steps when consecutive
+timestep embeddings are similar. MiniMax-H3 TeaCache is currently calibrated
+only for the FL2VA partition. In combined serving, FL2VA requests use TeaCache
+while Ref2VA requests run uncached; Ref2VA-only serving rejects TeaCache.
+
+TeaCache and Cache-DiT are mutually exclusive; pick one cache backend per server.
+
+The model-specific default and examples use `rel_l1_thresh=0.17`, which
+provided the best conservative speed/quality balance in the validated 107-frame
+T2VA workload. Lower values
+may produce few or no cache hits, while higher values can improve performance
+at the cost of output quality. Validate the threshold on representative
+prompts and generation settings before changing it.
+
+### Offline (Python API)
+
+Export the directory containing the `FL2VA` and `Ref2VA` subdirectories before
+running the Python example, for example `export MODEL_ROOT=/models/MiniMax-H3`.
+
+```python
+import os
+
+from vllm_omni import Omni
+from vllm_omni.inputs.data import OmniDiffusionSamplingParams
+
+omni = Omni(
+    model=os.path.join(os.environ["MODEL_ROOT"], "FL2VA"),
+    cache_backend="tea_cache",
+    cache_config={"rel_l1_thresh": 0.17},
+    trust_remote_code=True,
+    enable_cpu_offload=True,
+)
+outputs = omni.generate(
+    "A quiet cinematic night scene with matching ambient sound.",
+    OmniDiffusionSamplingParams(
+        height=256,
+        width=448,
+        num_frames=29,
+        fps=24,
+        num_inference_steps=50,
+        seed=42,
+        extra_args={
+            "task": "t2va",
+            "duration": 4.0,
+            "aspect_ratio": "16:9",
+            "flow_shift": 12.0,
+            "audio_flow_shift": 3.0,
+        },
+    ),
+)
+```
+
+### Online serving
+
+```bash
+vllm serve "${MODEL_ROOT}/FL2VA" \
+  --omni \
+  --trust-remote-code \
+  --cache-backend tea_cache \
+  --cache-config '{"rel_l1_thresh":0.17}'
+```
+
 ## Known limitations

+- TeaCache is calibrated for FL2VA only; Ref2VA requests run uncached.
 - Combined serving requires sibling `FL2VA` and `Ref2VA` directories, loads
  both task-specific DiTs, and loads shared components once from `FL2VA`.
 - H3 currently executes one generation request per diffusion batch.
@@ -770,6 +835,10 @@ reduction.
 - A U2 x Ring2 hybrid currently fails with an attention-mask length mismatch; use
  pure Ulysses.
 - 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
+  audio reference.
 - Pure Ulysses still replicates the full DiT on every rank, so smaller-memory GPUs
  cannot use `--usp N --tp 1` as a resident capacity path. Use DiT tensor parallelism
  or model-level CPU offload; text-encoder TP alone is not sufficient.
@@ -30,8 +30,9 @@ also provided.
  `hf_config.version="4.5"`):
  - Default single-GPU compatibility layout (auto-loaded):
    [`vllm_omni/deploy/minicpmo_4_5.yaml`](../../vllm_omni/deploy/minicpmo_4_5.yaml)
-  - 2-GPU and 3-GPU layouts:
+  - Recommended 2-GPU continuous-batching layout:
    [`vllm_omni/deploy/minicpmo_4_5_2gpu.yaml`](../../vllm_omni/deploy/minicpmo_4_5_2gpu.yaml),
+  - 3-GPU layout:
    [`vllm_omni/deploy/minicpmo_4_5_3gpu.yaml`](../../vllm_omni/deploy/minicpmo_4_5_3gpu.yaml)
  - 8x RTX 4090 layout:
    [`vllm_omni/deploy/minicpmo_4_5_8x4090.yaml`](../../vllm_omni/deploy/minicpmo_4_5_8x4090.yaml)
@@ -19,6 +19,11 @@ add files for this repo. Use one Markdown file per model family by default.
 > Individual pages resolve at `https://recipes.vllm.ai/<provider>/<model-id>`,
 > e.g. [`bosonai/higgs-audio-v3-tts-4b`](https://recipes.vllm.ai/bosonai/higgs-audio-v3-tts-4b).

+The `Recipe` column in the Supported Models documentation links to the
+published recipe when available and otherwise to the corresponding recipe in
+this directory. The recipe files remain the source of truth for commands and
+validation details.
+
 Example layout:

 ```text
@@ -50,6 +55,7 @@ recipes/
 | [`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 |
 | [`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 |
 | [`Robbyant/LingBot-Video.md`](./Robbyant/LingBot-Video.md) | Native dense and MoE text-to-video serving | Dense: 1x L20X; MoE: 1x L20X (~67.7 GiB peak) |
 | [`cosmos3/Cosmos3-Nano.md`](./cosmos3/Cosmos3-Nano.md) | Text-to-image, text-to-video, image-to-video, video-to-video generation, text to video with sound, action policy | 1x H200 141GB / B300 |
@@ -0,0 +1,282 @@
+# LongCat-Video-Avatar-1.5
+
+> Audio-driven avatar video generation (AT2V / AI2V)
+
+## Summary
+
+- Vendor: meituan-longcat
+- Model: `meituan-longcat/LongCat-Video-Avatar-1.5`
+- Task: Audio-to-video (AT2V) and audio-and-image-to-video (AI2V) avatar generation
+- Mode: Offline inference
+- Maintainer: Community
+
+## When to use this recipe
+
+Use this recipe to drive a talking or singing avatar from a speech clip, either
+from text alone (AT2V) or from a reference image (AI2V), with the native
+LongCat-Video-Avatar-1.5 pipeline in vLLM-Omni. It covers:
+
+1. **Single-speaker AT2V / AI2V** — one 93-frame segment from one audio track.
+2. **AVC continuation** — chained segments that cover a longer audio track.
+3. **Multi-speaker AI2V** — several audio tracks placed on a reference image
+   through speaker bounding boxes.
+
+The example saves an mp4 with the generated frames muxed with the input audio.
+
+## References
+
+- Upstream model card: <https://huggingface.co/meituan-longcat/LongCat-Video-Avatar-1.5>
+- Upstream repository and example assets: <https://github.com/meituan-longcat/LongCat-Video>
+- Related example under `examples/`: [`longcat_video/end2end.py`](../../examples/offline_inference/longcat_video/end2end.py)
+
+## Hardware Support
+
+## CUDA
+
+### 1× NVIDIA H100 (80 GB)
+
+#### Environment
+
+- OS: Linux
+- Python: 3.10+
+- Driver: NVIDIA driver with CUDA 12.x
+- vLLM version: Match the repository requirements for your checkout
+- vLLM-Omni version or commit: Use the commit you are deploying from
+
+#### Prerequisites
+
+Install the LongCat Avatar optional runtime dependencies:
+
+```bash
+pip install "vllm-omni[longcat-video-avatar]"
+```
+
+The example uses the `imageio_ffmpeg` executable installed by vLLM-Omni's
+`imageio[ffmpeg]` dependency.
+
+#### Command
+
+Audio-to-video (AT2V):
+
+```bash
+python examples/offline_inference/longcat_video/end2end.py \
+  --model meituan-longcat/LongCat-Video-Avatar-1.5 \
+  --stage at2v \
+  --audio /path/to/speech.wav \
+  --prompt "A person speaks calmly while facing the camera." \
+  --num-frames 93 \
+  --num-inference-steps 8 \
+  --output longcat_avatar_at2v.mp4
+```
+
+Audio-and-image-to-video (AI2V):
+
+```bash
+python examples/offline_inference/longcat_video/end2end.py \
+  --model meituan-longcat/LongCat-Video-Avatar-1.5 \
+  --stage ai2v \
+  --audio /path/to/speech.wav \
+  --image /path/to/reference.jpg \
+  --prompt "A person speaks calmly while facing the camera." \
+  --num-frames 93 \
+  --num-inference-steps 8 \
+  --output longcat_avatar_ai2v.mp4
+```
+
+#### Official Example Assets
+
+Download the official Avatar example assets:
+
+```bash
+mkdir -p longcat_avatar_assets/assets/avatar/single longcat_avatar_assets/assets/avatar/multi
+
+curl -L -o longcat_avatar_assets/assets/avatar/single/man.mp3 \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/single/man.mp3
+curl -L -o longcat_avatar_assets/assets/avatar/single/man.png \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/single/man.png
+curl -L -o longcat_avatar_assets/assets/avatar/single_example_1.json \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/single_example_1.json
+
+curl -L -o longcat_avatar_assets/assets/avatar/multi/sing.png \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/multi/sing.png
+curl -L -o longcat_avatar_assets/assets/avatar/multi/sing_man.WAV \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/multi/sing_man.WAV
+curl -L -o longcat_avatar_assets/assets/avatar/multi/sing_woman.WAV \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/multi/sing_woman.WAV
+curl -L -o longcat_avatar_assets/assets/avatar/multi/introduce.png \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/multi/introduce.png
+curl -L -o longcat_avatar_assets/assets/avatar/multi/introduce_man.mp3 \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/multi/introduce_man.mp3
+curl -L -o longcat_avatar_assets/assets/avatar/multi/introduce_woman.mp3 \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/multi/introduce_woman.mp3
+curl -L -o longcat_avatar_assets/assets/avatar/multi_example_1.json \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/multi_example_1.json
+curl -L -o longcat_avatar_assets/assets/avatar/multi_example_2.json \
+  https://raw.githubusercontent.com/meituan-longcat/LongCat-Video/main/assets/avatar/multi_example_2.json
+
+export LONGCAT_VIDEO_ASSET_DIR="$PWD/longcat_avatar_assets/assets/avatar"
+export AVATAR_MODEL=meituan-longcat/LongCat-Video-Avatar-1.5
+export PROMPT="$(python -c 'import json, os; print(json.load(open(os.environ["LONGCAT_VIDEO_ASSET_DIR"] + "/single_example_1.json"))["prompt"])')"
+```
+
+`--base-model-dir` points at a local LongCat-Video base checkpoint for the
+tokenizer, text encoder and VAE; omit it to download those components from
+Hugging Face.
+
+```bash
+export BASE_MODEL_DIR=/cache/longcat_models/LongCat-Video
+```
+
+#### Official Asset Smoke Test
+
+Audio-to-video with the official audio:
+
+```bash
+python examples/offline_inference/longcat_video/end2end.py \
+  --model "$AVATAR_MODEL" \
+  --base-model-dir "$BASE_MODEL_DIR" \
+  --stage at2v \
+  --audio "$LONGCAT_VIDEO_ASSET_DIR/single/man.mp3" \
+  --prompt "$PROMPT" \
+  --num-frames 93 \
+  --num-inference-steps 8 \
+  --output official_asset_longcat_avatar_at2v.mp4
+```
+
+Audio-and-image-to-video with the official audio and image:
+
+```bash
+python examples/offline_inference/longcat_video/end2end.py \
+  --model "$AVATAR_MODEL" \
+  --base-model-dir "$BASE_MODEL_DIR" \
+  --stage ai2v \
+  --audio "$LONGCAT_VIDEO_ASSET_DIR/single/man.mp3" \
+  --image "$LONGCAT_VIDEO_ASSET_DIR/single/man.png" \
+  --prompt "$PROMPT" \
+  --num-frames 93 \
+  --num-inference-steps 8 \
+  --output official_asset_longcat_avatar_ai2v.mp4
+```
+
+#### Official Single-Speaker AVC Continuation
+
+`--num-segments` chains AVC segments; `auto` covers the full input audio.
+
+```bash
+python examples/offline_inference/longcat_video/end2end.py \
+  --model "$AVATAR_MODEL" \
+  --base-model-dir "$BASE_MODEL_DIR" \
+  --input-json "$LONGCAT_VIDEO_ASSET_DIR/single_example_1.json" \
+  --num-frames 93 \
+  --num-cond-frames 13 \
+  --num-segments 5 \
+  --num-inference-steps 8 \
+  --output official_asset_longcat_avatar_single_example_1_ai2v_5seg.mp4
+```
+
+Replace `--num-segments 5` with `--num-segments auto` to run until the full
+official audio is covered. That is a slow archival validation path rather than a
+regular smoke test: about 24m28s on a Modal H100 for the full
+`single_example_1.json` audio with 8 inference steps.
+
+#### Official Multi-Speaker AI2V
+
+The multi-speaker cases use several audio tracks with speaker masks and optional
+bounding boxes. The command below generates the first 93-frame segment, matching
+the first call to the official `generate_ai2v()`.
+
+```bash
+python examples/offline_inference/longcat_video/end2end.py \
+  --model "$AVATAR_MODEL" \
+  --base-model-dir "$BASE_MODEL_DIR" \
+  --input-json "$LONGCAT_VIDEO_ASSET_DIR/multi_example_1.json" \
+  --num-frames 93 \
+  --num-inference-steps 8 \
+  --output official_asset_longcat_avatar_multi_example_1_ai2v.mp4
+```
+
+Use `multi_example_2.json` for the sequential-speaking case, and add
+`--num-cond-frames 13 --num-segments auto` to cover the full audio with AVC
+continuation.
+
+#### Verification
+
+```bash
+ffprobe -hide_banner longcat_avatar_ai2v.mp4
+```
+
+The output must report one video stream and one audio stream: the example muxes
+the generated frames with the input audio.
+
+#### Notes
+
+**Support matrix**
+
+| Mode | Speakers | Image required | AVC continuation |
+| --- | --- | --- | --- |
+| AT2V | Single | No | Yes |
+| AI2V | Single | Yes | Yes |
+| AT2V | Multi | Not supported | Not supported |
+| AI2V | Multi | Yes | Yes |
+
+Multi-speaker Avatar requires AI2V because speaker bounding boxes and masks are
+defined on a reference image; AT2V with several tracks fails with an explicit
+error. AVC continuation uses full-audio embeddings only when `--num-segments` is
+greater than `1` or set to `auto`; single-segment AT2V/AI2V keeps the shorter
+first-segment audio path.
+
+**Sampling parameters**
+
+| Parameter | LongCat pipeline value | Description |
+| --- | --- | --- |
+| `--num-frames` | `num_frames` | Frames generated per segment. The default is 93 frames, or about 3.72 seconds at 25 FPS. LongCat normalizes other values to the temporal-VAE-compatible `4k+1` form. |
+| `--num-inference-steps` | `steps` | Denoising steps. The distilled Avatar path uses 8 steps. |
+| `--fps` | `save_fps` | Output frame rate, and the rate used to align audio embeddings with video frames. The default is 25 FPS. |
+| `sampling_params.guidance_scale` | `text_guidance_scale` | Classifier-free guidance applied to the text prompt. |
+| `sampling_params.guidance_scale_2` | `audio_guidance_scale` | Classifier-free guidance applied to the speech audio. |
+
+With the distilled LoRA enabled (the default), the pipeline uses 8 inference
+steps and sets both guidance scales to `1.0`. With `--no-use-distill`, the
+example sets both to `4.0`. For AVC continuation every segment holds
+`num_frames` frames, but each segment after the first reuses `num_cond_frames`
+conditioning frames, so the unique frame count is
+`num_frames + (num_segments - 1) * (num_frames - num_cond_frames)`.
+
+**Key flags**
+
+- `--stage`: `at2v` for audio-to-video, `ai2v` for audio-and-image-to-video.
+- `--resolution`: `480p` or `720p`.
+- `--input-json`: official LongCat Avatar JSON case, used for the single-speaker
+  and multi-speaker AI2V examples.
+- `--num-segments`: number of AVC segments, or `auto` to cover the full audio.
+- `--num-cond-frames`: previous-frame conditioning window for AVC continuation.
+- `--use-int8` / `--no-use-int8`: load the official INT8 Avatar DiT weights by
+  default, or full precision weights.
+- `--use-distill` / `--no-use-distill`: enable the official distilled LoRA path
+  by default.
+- `--build-components-on-gpu`: build large Avatar components directly on GPU for
+  faster startup, at a higher peak VRAM cost. Disabled by default.
+- `--base-model-dir`: local LongCat-Video base model directory for the
+  tokenizer, text encoder and VAE components.
+
+**Memory usage**
+
+Measured on a Modal H100 running the official 93-frame AI2V example with the
+INT8 DiT, distilled LoRA and 8 inference steps:
+
+| Component build | CLI option | Peak GPU memory | Model loading time |
+| --- | --- | ---: | ---: |
+| CPU (default) | None | About 41.0 GiB | About 130s |
+| GPU | `--build-components-on-gpu` | About 56.8 GiB | About 33s |
+
+The reported peak covers the end-to-end run. Building components on CPU only
+reduces initialization memory; denoising still runs on GPU. These numbers are
+memory guidance rather than a performance benchmark.
+
+**Known limitations**
+
+- Single GPU only. Cache acceleration, sequence/CFG/tensor parallelism, HSDP,
+  CPU offload, VAE patch parallelism and quantization are not wired up for this
+  pipeline yet; see the diffusion feature matrix in
+  [`docs/user_guide/diffusion_features.md`](../../docs/user_guide/diffusion_features.md).
+- Offline inference only. Online serving is not supported yet.
@@ -6,13 +6,21 @@ Unit tests for patch.py
 """

 import asyncio
+import base64
 import json
+import time
+from types import SimpleNamespace

 import pytest
 from pytest_mock import MockerFixture
 from vllm.benchmarks.lib.endpoint_request_func import RequestFuncInput

-from vllm_omni.benchmarks.patch.patch import MixRequestFuncOutput, async_request_openai_chat_omni_completions
+from vllm_omni.benchmarks.patch.patch import (
+    MixRequestFuncOutput,
+    async_request_openai_chat_omni_completions,
+    async_request_openai_realtime_duplex,
+)
+from vllm_omni.experimental.fullduplex.client import RealtimeEventCollector

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

@@ -40,6 +48,156 @@ class MockResponse:
        pass


+@pytest.mark.asyncio
+async def test_seed_tts_realtime_duplex_exports_per_request_metrics(monkeypatch):
+    class FakeRealtimeClient:
+        last_instance = None
+
+        def __init__(self, url):
+            assert url == "ws://localhost:8000/v1/realtime?duplex=1"
+            self.events = RealtimeEventCollector()
+            self.configure_kwargs = None
+            self.sent = []
+            self.response_count = 0
+            self.ack_count = 0
+            type(self).last_instance = self
+
+        async def __aenter__(self):
+            return self
+
+        async def __aexit__(self, *_args):
+            return None
+
+        async def configure(self, model, **kwargs):
+            assert model == "openbmb/MiniCPM-o-4_5"
+            self.configure_kwargs = kwargs
+
+        async def send(self, event):
+            self.sent.append(event)
+            if event["type"] != "response.create":
+                return
+            self.response_count += 1
+            response_id = f"resp-{self.response_count}"
+            now = time.monotonic()
+            self.events.add(
+                {"type": "response.created", "response": {"id": response_id}},
+                received_at_s=now + 0.01,
+            )
+            self.events.add(
+                {
+                    "type": "response.audio_transcript.delta",
+                    "response_id": response_id,
+                    "delta": f"turn {self.response_count}",
+                },
+                received_at_s=now + 0.02,
+            )
+            self.events.add(
+                {
+                    "type": "response.audio.delta",
+                    "response_id": response_id,
+                    "delta": base64.b64encode(b"\x00\x00" * 2400).decode(),
+                    "sample_rate_hz": 24_000,
+                    "metadata": {"audio_duration_ms": 100},
+                },
+                received_at_s=now + 0.03,
+            )
+            self.events.add(
+                {"type": "response.done", "response": {"id": response_id}},
+                received_at_s=now + 0.04,
+            )
+
+        async def acknowledge_playback(self):
+            self.ack_count += 1
+
+        async def close_session(self, **_kwargs):
+            return None
+
+    monkeypatch.setattr(
+        "vllm_omni.benchmarks.patch.patch.RealtimeDuplexClient",
+        FakeRealtimeClient,
+    )
+    request_input = RequestFuncInput(
+        model="openbmb/MiniCPM-o-4_5",
+        model_name="openbmb/MiniCPM-o-4_5",
+        prompt="hello",
+        api_url="http://localhost:8000/v1/realtime",
+        prompt_len=1,
+        output_len=20,
+        logprobs=None,
+        multi_modal_content=None,
+        ignore_eos=False,
+        extra_body={"save_duplex_request_metrics": True},
+    )
+    request_input.seed_tts_speech_extra = {"ref_audio": "data:audio/wav;base64,AAAA"}
+    request_input.seed_tts_system_prompt = "Speak exactly."
+    request_input.seed_tts_turns = tuple(
+        SimpleNamespace(utterance_id=f"utt-{index}", target_text=f"text {index}") for index in range(4)
+    )
+
+    output = await async_request_openai_realtime_duplex(
+        request_input,
+        session=None,
+    )
+
+    client = FakeRealtimeClient.last_instance
+    assert client.configure_kwargs["native_duplex"] is False
+    assert client.configure_kwargs["extra_body"] == {
+        "ref_audio": "data:audio/wav;base64,AAAA",
+    }
+    assert client.sent == [
+        event
+        for index in range(4)
+        for event in (
+            {
+                "type": "conversation.item.create",
+                "item": {
+                    "type": "message",
+                    "role": "user",
+                    "content": [{"type": "input_text", "text": f"text {index}"}],
+                },
+            },
+            {"type": "response.create"},
+        )
+    ]
+    assert client.ack_count == 4
+    assert output.success is True
+    assert output.generated_text == "turn 1 turn 2 turn 3 turn 4"
+    assert output.audio_duration == pytest.approx(0.4)
+    assert output.ttft > 0
+    assert output.audio_ttfp > output.ttft
+    assert output.audio_rtf > 0
+    assert output.latency > 0
+    assert output.tts_turn_pcm_bytes == [b"\x00\x00" * 2400] * 4
+    assert output.tts_output_pcm_bytes == b"\x00\x00" * 9600
+    session_id = output.duplex_request_metrics[0]["session_id"]
+    assert len(output.duplex_request_metrics) == 4
+    for index, metrics in enumerate(output.duplex_request_metrics):
+        assert metrics == {
+            "session_id": session_id,
+            "request_index": index,
+            "utterance_id": f"utt-{index}",
+            "response_id": f"resp-{index + 1}",
+            "source": "client_monotonic_receive",
+            "measurement_origin": {
+                "ttft": "conversation.item.create client send to first non-empty text delta",
+                "ttfp": "conversation.item.create client send to first audio packet",
+                "rtf": "request-start-to-last-audio receive time divided by emitted audio duration",
+            },
+            "ttft_ms": pytest.approx(20.0, abs=2.0),
+            "ttfp_ms": pytest.approx(30.0, abs=2.0),
+            "rtf": pytest.approx(0.3, abs=0.03),
+            "audio_generation_ms": pytest.approx(30.0, abs=2.0),
+            "audio_duration_ms": 100.0,
+        }
+    assert output.duplex_session_metrics == {
+        "session_id": session_id,
+        "audio_turn_count": 4,
+        "mean_ttft_ms": pytest.approx(20.0, abs=2.0),
+        "mean_ttfp_ms": pytest.approx(30.0, abs=2.0),
+        "mean_rtf": pytest.approx(0.3, abs=0.03),
+    }
+
+
 def create_sse_chunk(data_dict):
    """Helper to create SSE formatted chunk"""
    return f"data: {json.dumps(data_dict)}\n\n".encode()
@@ -42,7 +42,10 @@ from benchmarks.accuracy.text_to_image.gbench import (
    summarize_generated_records as summarize_gebench_generated_records,
    summarize_gebench_results,
 )
-from tests.e2e.accuracy.qwen3_omni.qwen3_omni_acc_bench_core import seed_tts_bench_argv
+from tests.e2e.accuracy.qwen3_omni.qwen3_omni_acc_bench_core import (
+    build_serve_common_argv,
+    seed_tts_bench_argv,
+)
 from tests.e2e.accuracy.qwen3_omni.run_qwen_omni_acc_benchmark import sync_dataset_env_from_ns
 from vllm_omni.benchmarks.data_modules.seed_tts_dataset import resolve_seed_tts_root

@@ -57,6 +60,25 @@ def test_seed_tts_bench_argv_preserves_hf_repo_id_from_env(monkeypatch):
    assert argv[dataset_idx + 1] == "zhaochenyang20/seed-tts-eval"


+def test_build_serve_common_argv_accepts_realtime_seed_tts_route(tmp_path):
+    argv = build_serve_common_argv(
+        host="127.0.0.1",
+        port=8000,
+        model="openbmb/MiniCPM-o-4_5",
+        num_prompts=4,
+        max_concurrency=1,
+        num_warmups=0,
+        percentile_metrics="ttft,audio_ttfp,audio_rtf",
+        result_dir=tmp_path,
+        result_filename="result.json",
+        backend="openai-realtime-tts",
+        endpoint="/v1/realtime",
+    )
+
+    assert argv[argv.index("--backend") + 1] == "openai-realtime-tts"
+    assert argv[argv.index("--endpoint") + 1] == "/v1/realtime"
+
+
 def test_sync_dataset_env_preserves_seed_tts_hf_repo_id(monkeypatch):
    ns = argparse.Namespace(
        daily_omni_repo=None,
@@ -28,10 +28,13 @@ if _MODULE_NAME not in sys.modules:
    _spec.loader.exec_module(_mod)

 from vllm_omni.benchmarks.data_modules.seed_tts_dataset import (  # noqa: E402
+    SeedTTSDataset,
    SeedTTSDesignDataset,
    SeedTTSDesignSampleRequest,
+    SeedTTSSampleRequest,
    SeedTTSTextDataset,
    SeedTTSTextSampleRequest,
+    SeedTTSTurn,
 )

 # ---------------------------------------------------------------------------
@@ -70,6 +73,94 @@ def mock_tokenizer(mocker):
 # ---------------------------------------------------------------------------


+def test_seed_tts_dataset_groups_four_turns_with_one_reference(seed_tts_root, mock_tokenizer):
+    ds = SeedTTSDataset(
+        dataset_path=str(seed_tts_root),
+        random_seed=0,
+        locale="en",
+        disable_shuffle=True,
+    )
+
+    requests = ds.sample(
+        mock_tokenizer,
+        num_requests=1,
+        turns_per_session=4,
+        no_oversample=True,
+    )
+
+    assert len(requests) == 1
+    request = requests[0]
+    assert [turn.utterance_id for turn in request.seed_tts_turns] == [
+        "utt000",
+        "utt001",
+        "utt002",
+        "utt003",
+    ]
+    assert [turn.target_text for turn in request.seed_tts_turns] == [
+        "target text 0",
+        "target text 1",
+        "target text 2",
+        "target text 3",
+    ]
+    assert request.seed_tts_speech_extra["ref_text"] == "ref text 0"
+    assert request.seed_tts_ref_wav_path.endswith("utt000.wav")
+
+
+def test_seed_tts_eval_expands_grouped_pcm_by_turn():
+    from vllm_omni.benchmarks.data_modules.seed_tts_eval import (
+        _expand_seed_tts_turn_outputs,
+    )
+
+    request = SeedTTSSampleRequest(
+        prompt="target 0",
+        prompt_len=4,
+        expected_output_len=100,
+        multi_modal_data=None,
+        request_id="session-0",
+        seed_tts_turns=tuple(
+            SeedTTSTurn(utterance_id=f"utt-{index}", target_text=f"target {index}") for index in range(4)
+        ),
+    )
+    output = types.SimpleNamespace(
+        success=True,
+        tts_turn_pcm_bytes=[bytes([index]) for index in range(4)],
+    )
+
+    requests, outputs = _expand_seed_tts_turn_outputs([request], [output])
+
+    assert [item.prompt for item in requests] == [f"target {index}" for index in range(4)]
+    assert [item.seed_tts_utterance_id for item in requests] == [f"utt-{index}" for index in range(4)]
+    assert [item.tts_output_pcm_bytes for item in outputs] == [bytes([index]) for index in range(4)]
+
+
+def test_seed_tts_eval_keeps_session_pcm_for_single_turn_chat():
+    """Chat-omni Seed-TTS sets session PCM only; expand must not wipe it."""
+    from vllm_omni.benchmarks.data_modules.seed_tts_eval import (
+        _expand_seed_tts_turn_outputs,
+    )
+
+    request = SeedTTSSampleRequest(
+        prompt="target 0",
+        prompt_len=4,
+        expected_output_len=100,
+        multi_modal_data=None,
+        request_id="session-0",
+        seed_tts_turns=(SeedTTSTurn(utterance_id="utt-0", target_text="target 0"),),
+    )
+    pcm = b"\x01\x02\x03\x04"
+    output = types.SimpleNamespace(
+        success=True,
+        tts_output_pcm_bytes=pcm,
+        tts_turn_pcm_bytes=None,
+    )
+
+    requests, outputs = _expand_seed_tts_turn_outputs([request], [output])
+
+    assert len(requests) == 1
+    assert requests[0].prompt == "target 0"
+    assert outputs[0].tts_output_pcm_bytes == pcm
+
+
 def test_seed_tts_text_dataset_omits_ref_audio(seed_tts_root, mock_tokenizer):
    ds = SeedTTSTextDataset(
        dataset_path=str(seed_tts_root),
@@ -2206,6 +2206,24 @@ class TestBaseConfigInheritance:
 class TestPlatformOverrides:
    """Test platform-specific deploy config overrides."""

+    def test_qwen3_omni_musa_clears_parent_quantization_for_audio_stages(self):
+        pipeline = resolve_pipeline_config(
+            "qwen3_omni_moe",
+            Q3_OMNI_ALL_STAGES_HF_CONFIG,
+        )
+        assert isinstance(pipeline, PipelineConfig)
+        deploy_path = Path(get_deploy_config_path("qwen3_omni_moe.yaml"))
+
+        base_stages = merge_pipeline_deploy(pipeline, load_deploy_config(deploy_path))
+        assert "hf_overrides" not in base_stages[1].yaml_engine_args
+        assert "hf_overrides" not in base_stages[2].yaml_engine_args
+
+        musa = _apply_platform_overrides(load_deploy_config(deploy_path), platform="musa")
+        musa_stages = merge_pipeline_deploy(pipeline, musa)
+        expected = {"quantization_config": None}
+        assert musa_stages[1].yaml_engine_args["hf_overrides"] == expected
+        assert musa_stages[2].yaml_engine_args["hf_overrides"] == expected
+
    def test_qwen3_tts_rocm_disables_code2wav_outer_cudagraph(self):
        deploy_path = Path(get_deploy_config_path("qwen3_tts.yaml"))

@@ -2241,7 +2259,8 @@ class TestPlatformOverrides:
                load_deploy_config(Path(get_deploy_config_path(filename))), platform="cuda"
            )
            replica_stages = merge_pipeline_deploy(pipeline, replica)
-            assert replica_stages[1].yaml_engine_args["kv_cache_memory_bytes"] is None
+            # Explicit null clears the inherited single-GPU 2 GiB CUDA cap.
+            assert replica_stages[1].yaml_engine_args.get("kv_cache_memory_bytes") is None

    def test_npu_overrides(self):
        deploy_path = Path(get_deploy_config_path("qwen3_omni_moe.yaml"))
@@ -39,6 +39,7 @@ from vllm_omni.config.stage_config import (
    load_deploy_config,
    merge_pipeline_deploy,
 )
+from vllm_omni.diffusion.diffusion_kv.config import DiffusionKVCacheMode
 from vllm_omni.engine.stage_init_utils import build_legacy_engine_args_dict

 pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
@@ -473,6 +474,8 @@ def test_sub_config_fields_match_structured_scopes():
        "moe_backend",
        "hf_overrides",
        "limit_mm_per_prompt",
+        "interleave_mm_strings",
+        "media_io_kwargs",
        "active_stream_window",
        "duplex_max_sessions",
        "enable_sleep_mode",
@@ -966,20 +969,24 @@ def test_diffusion_config_from_kwargs_reuses_legacy_normalization(monkeypatch):

    cfg = omni_config_module._DiffusionConfigProjection.from_kwargs(
        diffusion_attention_backend="flash_attn",
+        fa_deterministic=True,
        kv_cache_dtype="fp8",
        kv_cache_skip_steps="0-1",
        kv_cache_skip_layers=[2],
        static_lora_scale=0.25,
+        diffusion_kv_mode="paged_scheduler",
        diffusers_load_kwargs=None,
        diffusers_call_kwargs=None,
    )

    assert cfg.diffusion_attention_config.default.backend == "flash_attn"
+    assert cfg.fa_deterministic is True
    assert cfg.diffusion_kv_cache_dtype == "fp8"
    assert cfg.diffusion_kv_cache_skip_step_indices == {0, 1}
    assert cfg.diffusion_kv_cache_skip_layer_indices == {2}
    assert cfg.lora_scale == 0.25
    assert cfg.cache_backend == "tea_cache"
+    assert cfg.diffusion_kv_mode is DiffusionKVCacheMode.PAGED_SCHEDULER
    assert cfg.diffusers_load_kwargs == {}
    assert cfg.diffusers_call_kwargs == {}

@@ -994,6 +1001,8 @@ def test_from_pipeline_config_normalizes_diffusion_config_aliases_from_engine_ar
                "stages:",
                "  - stage_id: 0",
                "    diffusion_attention_backend: flash_attn",
+                "    fa_deterministic: true",
+                "    diffusion_kv_mode: paged_scheduler",
            ]
        )
    )
@@ -1005,6 +1014,29 @@ def test_from_pipeline_config_normalizes_diffusion_config_aliases_from_engine_ar

    assert isinstance(stage, VllmOmniDiffusionStageConfig)
    assert stage.diffusion_config.diffusion_attention_config.default.backend == "flash_attn"
+    assert stage.diffusion_config.fa_deterministic is True
+    assert stage.diffusion_config.diffusion_kv_mode is DiffusionKVCacheMode.PAGED_SCHEDULER
+
+
+def test_from_pipeline_config_rejects_reserved_diffusion_kv_mode(tmp_path):
+    deploy_path = tmp_path / "dreamzero_reserved_diffusion_kv.yaml"
+    deploy_path.write_text(
+        "\n".join(
+            [
+                "pipeline: dreamzero",
+                "async_chunk: false",
+                "stages:",
+                "  - stage_id: 0",
+                "    diffusion_kv_mode: paged_worker_local",
+            ]
+        )
+    )
+
+    with pytest.raises(ValueError, match="reserved but not implemented"):
+        _from_pipeline_key(
+            "dreamzero",
+            deploy_config_path=str(deploy_path),
+        )


 def test_diffusion_config_field_classification_covers_current_fields():
@@ -1025,6 +1057,7 @@ def test_diffusion_config_field_classification_covers_current_fields():
        "enable_session_state_manager",
        "prompt_embed_cache_size",
        "diffusion_kv_cache_dtype",
+        "diffusion_kv_mode",
    } <= omni_config_module._DIFFUSION_ONLY_CONFIG_FIELDS
    assert {
        "revision",
@@ -92,6 +92,9 @@ class _Request:
        self.num_computed_tokens = 0
        self.num_in_flight_tokens = 0
        self.num_output_placeholders = 0
+        # vLLM 0.27 (a0c092ee72): Request gained num_stale_output_tokens to
+        # track in-flight outputs discarded at preemption/streaming-stop.
+        self.num_stale_output_tokens = 0
        self.has_encoder_inputs = False
        self.pooling_params = None
        self.resumable = True
@@ -0,0 +1,223 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+"""Investigation tests for the num_stale_output_tokens drain (59b9e719).
+
+The seed-TTS WER regression appeared with the vLLM 0.27 rebase (nightly
+2946 -> 2949: mean WER 0.2695 -> 0.3936, median 0.0 in both), whose prime
+suspect is the reimplemented async-token discard: vLLM 0.26 drained
+``async_tokens_to_discard`` one arriving frame at a time and suppressed
+only the token append; the 0.27 adaptation seeds
+``num_stale_output_tokens`` with the discarded *placeholder count* and
+drains it by each arriving frame's ``num_tokens_scheduled``, dropping the
+whole per-request update. Whenever seed and arrivals disagree, the counter
+outlives the old segment and the drain consumes *valid* frames of the new
+segment — a hole in the talker's codec stream (garbled audio, text/audio
+mismatch) — or underflows its own assert on the new segment's prefill
+frame (engine-core death mid-step).
+
+The fix seeds the counter from ``num_in_flight_tokens`` — the same
+scheduled-token units the drain subtracts — so every pre-discard frame
+drains exactly its own contribution and new-segment frames can never be
+swallowed. These tests pin that contract one frame at a time; the
+delivered / swallowed observable is whether ``_update_request_with_output``
+runs for the frame.
+"""
+
+from __future__ import annotations
+
+from types import SimpleNamespace
+from unittest.mock import MagicMock
+
+import pytest
+
+# Imports must run in this order: vllm_omni applies patches to vllm.v1.request
+# before Request / StreamingUpdate are bound in this module.
+# isort: off
+import vllm_omni  # noqa: F401 - import for side effects (patch vLLM)
+from vllm.sampling_params import SamplingParams
+from vllm.v1.core.sched.output import SchedulerOutput
+from vllm.v1.outputs import ModelRunnerOutput
+from vllm.v1.request import Request, RequestStatus, StreamingUpdate
+from vllm_omni.core.sched.omni_ar_scheduler import OmniARScheduler
+
+# isort: on
+
+pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
+
+
+def _make_session() -> Request:
+    session = Request(
+        request_id="req-stale-drain-test",
+        prompt_token_ids=[1, 2, 3],
+        sampling_params=SamplingParams(max_tokens=32),
+        pooling_params=None,
+        arrival_time=100.0,
+        block_hasher=None,
+    )
+    session.status = RequestStatus.RUNNING
+    if not hasattr(session, "num_stale_output_tokens"):
+        # vLLM 0.27 defines this Request field; 0.26 (local dev) does not.
+        session.num_stale_output_tokens = 0
+    if not hasattr(session, "num_in_flight_tokens"):
+        session.num_in_flight_tokens = 0
+    return session
+
+
+def _replace_streaming_session(session: Request) -> None:
+    """Run the real stage-0 streaming replacement (the discard/seed site)."""
+    sched = OmniARScheduler.__new__(OmniARScheduler)
+    sched._new_prompt_len_snapshot = {}
+    sched.vllm_config = SimpleNamespace(model_config=SimpleNamespace(stage_id=0))
+    sched.num_waiting_for_streaming_input = 0
+    sched.log_stats = False
+    sched.chunk_transfer_adapter = None
+    sched.skipped_waiting = set()
+    session.status = RequestStatus.WAITING_FOR_STREAMING_REQ
+    update = StreamingUpdate(
+        mm_features=None,
+        prompt_token_ids=[10, 20],
+        max_tokens=32,
+        arrival_time=200.0,
+        sampling_params=SamplingParams(max_tokens=16),
+    )
+    sched._update_request_as_session(session, update)
+    session.status = RequestStatus.RUNNING
+
+
+def _make_drain_sched(session: Request) -> MagicMock:
+    sched = MagicMock()
+    sched.requests = {session.request_id: session}
+    sched.perf_metrics = None
+    sched.structured_output_manager.should_advance.return_value = False
+    sched._update_request_with_output.return_value = ([42], False)
+    sched._process_kv_transfer_trigger.return_value = False
+    sched.chunk_transfer_adapter = MagicMock()
+    sched.running = [session]
+    sched.waiting_for_transfer_free = set()
+    sched.transfer_triggered_requests = set()
+    sched.active_kv_transfers = set()
+    sched.pending_stop_after_extraction = set()
+    sched.connector = None
+    sched.kv_cache_manager.take_events.return_value = None
+    sched.finished_req_ids_dict = {}
+    sched.make_stats.return_value = None
+    return sched
+
+
+def _run_step(sched: MagicMock, session: Request, *, num_scheduled: int, token: int) -> bool:
+    """Feed one model-runner frame through update_from_output.
+
+    Returns True when the frame's tokens were delivered (the append ran),
+    False when the frame was swallowed by the stale drain.
+    """
+    scheduler_output = MagicMock(spec=SchedulerOutput)
+    scheduler_output.num_scheduled_tokens = {session.request_id: num_scheduled}
+    scheduler_output.scheduled_spec_decode_tokens = {}
+    scheduler_output.num_invalid_spec_tokens = 0
+
+    model_runner_output = MagicMock(spec=ModelRunnerOutput)
+    model_runner_output.sampled_token_ids = [[token]]
+    model_runner_output.logprobs = None
+    model_runner_output.prompt_logprobs_dict = {}
+    model_runner_output.pooler_output = None
+    model_runner_output.num_nans_in_logits = None
+    model_runner_output.kv_connector_output = None
+    model_runner_output.cudagraph_stats = None
+    model_runner_output.req_id_to_index = {session.request_id: 0}
+    model_runner_output.routed_experts = None
+    model_runner_output.inter_stage_outputs = None
+
+    sched._update_request_with_output.reset_mock()
+    OmniARScheduler.update_from_output(sched, scheduler_output, model_runner_output)
+    return sched._update_request_with_output.called
+
+
+def test_exact_drain_delivers_new_segment_frame() -> None:
+    """One in-flight decode frame at replacement: its late output drains the
+    counter exactly and the new segment's first frame is delivered."""
+    session = _make_session()
+    session.append_output_token_ids([7, 8, 9])
+    session.num_computed_tokens = 6
+    session.num_output_placeholders = 1
+    session.num_in_flight_tokens = 1
+
+    _replace_streaming_session(session)
+    assert session.num_stale_output_tokens == 1
+
+    sched = _make_drain_sched(session)
+    assert _run_step(sched, session, num_scheduled=1, token=42) is False  # late frame dropped
+    assert session.num_stale_output_tokens == 0
+    assert _run_step(sched, session, num_scheduled=1, token=43) is True  # new segment survives
+
+
+def test_two_in_flight_frames_drain_without_swallowing() -> None:
+    """Two in-flight decode frames: both late frames drop, the new segment's
+    first valid frame is delivered. A placeholder-based seed swallowed it
+    whenever fewer frames arrived than placeholders counted (the driver of
+    the seed-TTS WER regression, nightly 2946 -> 2949: 0.2695 -> 0.3936)."""
+    session = _make_session()
+    session.append_output_token_ids([7, 8, 9])
+    session.num_computed_tokens = 7
+    session.num_output_placeholders = 2
+    session.num_in_flight_tokens = 2
+
+    _replace_streaming_session(session)
+    assert session.num_stale_output_tokens == 2
+
+    sched = _make_drain_sched(session)
+    assert _run_step(sched, session, num_scheduled=1, token=42) is False  # late frame 1
+    assert _run_step(sched, session, num_scheduled=1, token=44) is False  # late frame 2
+    assert session.num_stale_output_tokens == 0
+
+    delivered = _run_step(sched, session, num_scheduled=1, token=43)  # new segment, frame 1
+    assert delivered is True, (
+        "valid new-segment frame was swallowed by leftover stale count "
+        f"(num_stale_output_tokens left at {session.num_stale_output_tokens})"
+    )
+
+
+def test_spec_placeholder_asymmetry_does_not_swallow_valid_frame() -> None:
+    """Placeholder counts can exceed scheduled counts (e.g. a draft
+    invalidated before execution): 1 in-flight MTP frame scheduled 2 tokens
+    but contributed 3 placeholders. Seeding from num_in_flight_tokens keeps
+    seed == drain, so the new segment's valid frame is delivered."""
+    session = _make_session()
+    session.append_output_token_ids([7])
+    session.num_computed_tokens = 7
+    session.num_output_placeholders = 3  # 1 sampled + 2 spec placeholders
+    session.num_in_flight_tokens = 2  # the frame actually scheduled 1 + 1 tokens
+
+    _replace_streaming_session(session)
+    assert session.num_stale_output_tokens == 2
+
+    sched = _make_drain_sched(session)
+    assert _run_step(sched, session, num_scheduled=2, token=42) is False  # late MTP frame
+    assert session.num_stale_output_tokens == 0
+    delivered = _run_step(sched, session, num_scheduled=1, token=43)  # new segment, frame 1
+    assert delivered is True, (
+        "valid new-segment frame swallowed after spec placeholder asymmetry "
+        f"(num_stale_output_tokens left at {session.num_stale_output_tokens})"
+    )
+
+
+def test_in_flight_prefill_chunk_drains_exactly_without_underflow() -> None:
+    """An in-flight prefill chunk carries no placeholders but 5 scheduled
+    tokens. The placeholder-based seed missed it entirely (its late output
+    was appended to the rolled-back session) and any leftover counter
+    underflowed 'assert num_stale_output_tokens >= 0' on the chunk's arrival,
+    killing the engine-core step. Seeding from num_in_flight_tokens drops the
+    late chunk and drains to exactly zero."""
+    session = _make_session()
+    session.append_output_token_ids([7, 8, 9])
+    session.num_computed_tokens = 7
+    session.num_output_placeholders = 0  # prefill chunk: no placeholders
+    session.num_in_flight_tokens = 5  # the chunk's scheduled tokens
+
+    _replace_streaming_session(session)
+    assert session.num_stale_output_tokens == 5
+
+    sched = _make_drain_sched(session)
+    assert _run_step(sched, session, num_scheduled=5, token=42) is False  # late prefill chunk dropped
+    assert session.num_stale_output_tokens == 0
+    assert _run_step(sched, session, num_scheduled=1, token=43) is True  # new segment survives
@@ -752,3 +752,12 @@ def pytest_addoption(parser: pytest.Parser) -> None:
        default=None,
        help=("Path to benchmark config JSON. Example: --test-config-file tests/dfx/perf/tests/test_tts.json"),
    )
+    parser.addoption(
+        "--assert-baseline",
+        action="store_true",
+        default=False,
+        help=(
+            "When set, omni/diffusion perf runners compare metrics against the baseline block in the JSON config "
+            "(default: off)."
+        ),
+    )
@@ -157,8 +157,19 @@ def benchmark_params(request):
    }


-def assert_result(result, num_prompt) -> None:
+def assert_result(result, params, num_prompt) -> None:
    assert result["completed"] == num_prompt, "Request failures exist"
+    expected_audio_turns = params.get("expected_duplex_audio_turns_per_session")
+    if expected_audio_turns is not None:
+        session_metrics = result.get("duplex_session_metrics")
+        assert isinstance(session_metrics, list), "Duplex session metrics are missing"
+        assert len(session_metrics) == num_prompt, (
+            f"Expected {num_prompt} duplex session metric rows, got {len(session_metrics)}"
+        )
+        assert all(
+            isinstance(metric, dict) and metric.get("audio_turn_count") == expected_audio_turns
+            for metric in session_metrics
+        ), f"Not every duplex session emitted {expected_audio_turns} audio turns"


@pytest.mark.benchmark
@@ -212,6 +223,7 @@ def test_performance_benchmark(omni_server, benchmark_params):
        "enabled",
        "eval_phase",
        "trust_remote_code",
+        "expected_duplex_audio_turns_per_session",
    }

    for key, value in params.items():
@@ -251,7 +263,7 @@ def test_performance_benchmark(omni_server, benchmark_params):
            random_output_len=params.get("random_output_len"),
            resource_label=resource_label,
        )
-        assert_result(result, num_prompt)
+        assert_result(result, params, num_prompt)

    # concurrency test
    for sweep_index, (concurrency, num_prompt) in enumerate(zip(max_concurrency_list, num_prompt_list)):
@@ -268,4 +280,4 @@ def test_performance_benchmark(omni_server, benchmark_params):
            random_output_len=params.get("random_output_len"),
            resource_label=resource_label,
        )
-        assert_result(result, num_prompt)
+        assert_result(result, params, num_prompt)
@@ -0,0 +1,87 @@
+[
+  {
+    "test_name": "test_minicpmo_4_5",
+    "mark": [
+      {
+        "hardware_marks": {
+          "res": {
+            "cuda": "H100",
+            "npu": "A3"
+          },
+          "num_cards": 1
+        }
+      },
+      "full_model",
+      "omni"
+    ],
+    "server_params": {
+      "model": "openbmb/MiniCPM-o-4_5",
+      "extra_cli_args": [
+        "--deploy-config",
+        "vllm_omni/deploy/minicpmo_4_5.yaml",
+        "--trust-remote-code"
+      ]
+    },
+    "benchmark_params": [
+      {
+        "dataset_name": "seed-tts",
+        "dataset_path": "zhaochenyang20/seed-tts-eval",
+        "backend": "openai-chat-omni",
+        "endpoint": "/v1/chat/completions",
+        "num_prompts": [
+          32,
+          64,
+          128
+        ],
+        "max_concurrency": [
+          1,
+          4,
+          8
+        ],
+        "no_oversample": true,
+        "trust_remote_code": true,
+        "seed_tts_locale": "en",
+        "extra_body": {
+          "modalities": [
+            "text",
+            "audio"
+          ],
+          "chat_template_kwargs": {
+            "enable_thinking": false,
+            "use_tts_template": true
+          }
+        },
+        "percentile-metrics": "ttft,e2el,audio_ttfp,audio_rtf,audio_duration",
+        "baseline": {
+          "H100": {
+            "request_throughput": [
+              0.7803,
+              1.1008,
+              1.2027
+            ],
+            "mean_ttft_ms": [
+              166.9764,
+              307.7133,
+              323.584
+            ],
+            "mean_e2el_ms": [
+              1361.1682,
+              3800.6332,
+              6961.8636
+            ],
+            "mean_audio_ttfp_ms": [
+              635.8099,
+              1615.6591,
+              2888.9831
+            ],
+            "mean_audio_rtf": [
+              0.3151,
+              0.8568,
+              1.5425
+            ]
+          }
+        }
+      }
+    ]
+  }
+]
@@ -0,0 +1,68 @@
+[
+  {
+    "test_name": "test_minicpmo_4_5_duplex_seed_tts",
+    "mark": [
+      {
+        "hardware_marks": {
+          "res": {
+            "cuda": "H100",
+            "npu": "A3"
+          },
+          "num_cards": 1
+        }
+      },
+      "full_model",
+      "omni"
+    ],
+    "server_params": {
+      "model": "openbmb/MiniCPM-o-4_5",
+      "extra_cli_args": [
+        "--deploy-config",
+        "vllm_omni/deploy/minicpmo_4_5_duplex.yaml",
+        "--trust-remote-code"
+      ]
+    },
+    "benchmark_params": [
+      {
+        "dataset_name": "seed-tts",
+        "dataset_path": "zhaochenyang20/seed-tts-eval",
+        "backend": "openai-realtime-tts",
+        "endpoint": "/v1/realtime",
+        "num_prompts": [
+          1
+        ],
+        "max_concurrency": [
+          1
+        ],
+        "no_oversample": true,
+        "trust_remote_code": true,
+        "seed_tts_locale": "en",
+        "seed_tts_turns_per_session": 4,
+        "expected_duplex_audio_turns_per_session": 4,
+        "extra_body": {
+          "save_duplex_request_metrics": true
+        },
+        "percentile-metrics": "ttft,e2el,audio_ttfp,audio_rtf,audio_duration",
+        "baseline": {
+          "H100": {
+            "request_throughput": [
+              0.1594
+            ],
+            "mean_ttft_ms": [
+              274.164
+            ],
+            "mean_e2el_ms": [
+              6250.7696
+            ],
+            "mean_audio_ttfp_ms": [
+              994.422
+            ],
+            "mean_audio_rtf": [
+              0.46163
+            ]
+          }
+        }
+      }
+    ]
+  }
+]
@@ -485,3 +485,30 @@ def test_diffusion_build_run_params_resolves_baseline_per_sweep(tmp_path, monkey
    assert sweeps[2]["params"]["baseline"]["H100"]["throughput_qps"] == 0.3
    assert "A3" in sweeps[2]["params"]["baseline"]
    assert not isinstance(sweeps[2]["params"]["baseline"]["H100"]["throughput_qps"], list)
+
+
+def test_omni_duplex_expected_audio_turns_accepts_complete_session():
+    from tests.dfx.perf.scripts.run_benchmark import assert_result
+
+    assert_result(
+        {
+            "completed": 1,
+            "duplex_session_metrics": [{"audio_turn_count": 4}],
+        },
+        {"expected_duplex_audio_turns_per_session": 4},
+        1,
+    )
+
+
+def test_omni_duplex_expected_audio_turns_rejects_incomplete_session():
+    from tests.dfx.perf.scripts.run_benchmark import assert_result
+
+    with pytest.raises(AssertionError, match="emitted 4 audio turns"):
+        assert_result(
+            {
+                "completed": 1,
+                "duplex_session_metrics": [{"audio_turn_count": 3}],
+            },
+            {"expected_duplex_audio_turns_per_session": 4},
+            1,
+        )
@@ -43,9 +43,9 @@
        ],
        "baseline": {
          "H100": {
-            "throughput_qps": 0.0434,
-            "latency_mean": 23.0654,
-            "peak_memory_mb_mean": 76360.0
+            "throughput_qps": 0.0361,
+            "latency_mean": 27.6206,
+            "peak_memory_mb_mean": 80548.0
          }
        }
      }
@@ -100,9 +100,9 @@
        ],
        "baseline": {
          "H100": {
-            "throughput_qps": 0.0606,
-            "latency_mean": 16.4928,
-            "peak_memory_mb_mean": 47164.8
+            "throughput_qps": 0.05263,
+            "latency_mean": 18.9415,
+            "peak_memory_mb_mean": 50053.7
          }
        }
      },
@@ -127,9 +127,9 @@
        ],
        "baseline": {
          "H100": {
-            "throughput_qps": 0.0124,
-            "latency_mean": 80.7784,
-            "peak_memory_mb_mean": 55374.3429
+            "throughput_qps": 0.009595,
+            "latency_mean": 104.1058,
+            "peak_memory_mb_mean": 59718.0
          }
        }
      }
@@ -50,12 +50,12 @@ def _prompt() -> str:
    return "What is the capital of China? Answer in 20 words."


-@hardware_test(res={"cuda": "H100", "npu": "A2"}, num_cards=1)
+@hardware_test(res={"cuda": "H100", "npu": "A3"}, num_cards=1)
@pytest.mark.parametrize("omni_server", _MINICPMO_SERVER, indirect=True)
 def test_invalid_audio_format_rejected(omni_server: OmniServerParams, openai_client: OpenAIClientHandler) -> None:
    """
    Test that invalid audio format is properly rejected.
-    Deploy Setting: default 2GPU
+    Deploy Setting: default single GPU
    Input Modal: text + invalid audio format
    Output Modal: error
    """
@@ -315,11 +315,32 @@ def _assert_post_fault_health_terminal(host: str, port: int, *, scenario: str) -
    pytest.fail(f"[{scenario} health] no terminal post-fault health observed: {last_observation}")


+def _assert_server_alive_after_fault(host: str, port: int, *, scenario: str) -> None:
+    """#4285: a single-stage engine death must NOT take down the API server.
+
+    The process must stay up and keep answering /health (503 is acceptable —
+    errored but listening). A connection-refused means the server process
+    exited, which is the bug this fix closes.
+    """
+    deadline = time.monotonic() + 15.0
+    responded = False
+    while time.monotonic() < deadline:
+        try:
+            status, _ = get_health_raw(host, port, timeout_sec=5)
+            responded = True
+            assert status in (200, 503), f"[{scenario} alive] unexpected /health status={status}"
+        except Exception as exc:  # noqa: BLE001
+            if _looks_like_server_unreachable(exc):
+                pytest.fail(f"[{scenario} alive] server exited after single-stage OOM (#4285): {exc!r}")
+            raise
+        time.sleep(0.5)
+    assert responded, f"[{scenario} alive] /health never responded after fault"
+
+
 QWEN_PARAMS = create_reliability_omni_server_params(RELIABILITY_SCENARIOS, DEPLOY_CONFIGS_DIR)


@pytest.mark.slow
-@pytest.mark.skip(reason="issue#4285")
@hardware_test(res={"cuda": "H100"}, num_cards=2)
@pytest.mark.parametrize("omni_server_function", QWEN_PARAMS, indirect=True)
 def test_reliability_fault_gpu_oom_error_contract_consistent_chat_speech(
@@ -386,9 +407,11 @@ def test_reliability_fault_gpu_oom_error_contract_consistent_chat_speech(
    assert "code" in chat_error, f"chat error lacks code field: {chat_error!r}"
    assert "code" in speech_error, f"speech error lacks code field: {speech_error!r}"

+    # #4285: a single dead stage must not take the whole API server down.
+    _assert_server_alive_after_fault(host, port, scenario="oom_contract")
+

@pytest.mark.slow
-@pytest.mark.skip(reason="issue#4285")
@hardware_test(res={"cuda": "H100"}, num_cards=2)
@pytest.mark.parametrize("omni_server_function", QWEN_PARAMS, indirect=True)
 def test_reliability_fault_gpu_oom_chat_large_payload_failure(omni_server_function, openai_client_function) -> None:
@@ -431,7 +454,6 @@ def test_reliability_fault_gpu_oom_chat_large_payload_failure(omni_server_functi


@pytest.mark.slow
-@pytest.mark.skip(reason="issue#4285")
@hardware_test(res={"cuda": "H100"}, num_cards=2)
@pytest.mark.parametrize("omni_server_function", QWEN_PARAMS, indirect=True)
 def test_reliability_fault_gpu_oom_concurrent_pressure_failure(omni_server_function, openai_client_function) -> None:
@@ -774,7 +796,11 @@ def test_reliability_fault_process_kill_tree_with_load_fast_fail_and_cleanup(


@pytest.mark.slow
-@pytest.mark.skip(reason="issue#4285")
+@pytest.mark.skip(
+    reason="needs request-scoped OOM recovery: under transient pressure the stage "
+    "EngineCore still dies and is evicted (not respawned), so the engine does not "
+    "recover after the hog stops. Out of scope for the #4285 fault-isolation fix."
+)
@hardware_test(res={"cuda": "H100"}, num_cards=2)
@pytest.mark.parametrize("omni_server_function", QWEN_PARAMS, indirect=True)
 def test_reliability_fault_gpu_oom_state_converges_after_fault_removed(
@@ -9,6 +9,10 @@ This script tests two main scenarios:
 2. Case 2: Comparing FlashAttention and SDPA backends for batch_size=2 with padding
 """

+import sys
+from types import SimpleNamespace
+from unittest.mock import Mock
+
 import pytest
 import torch

@@ -371,6 +375,264 @@ def test_packed_varlen_metadata_must_be_complete(monkeypatch):
        impl.forward_cuda(query, query, query, metadata)


+# ---------------------------------------------------------------------------
+# NPU mask-free packed attention (PR #5891).
+#
+# These tests cover the backend-side NPU branches that ``forward_fa_npu`` adds:
+#   - boundary resolution in ``_resolve_packed_seq_npu`` (pure Python, no device
+#     sync, no MindIE-SD import);
+#   - env dispatch in ``forward_fa_npu`` (which op a packed forward takes);
+#   - the laser input pre-scaling math in ``_forward_prefix_kv_slice_npu``.
+#
+# They run on CPU without a real NPU by injecting a fake ``mindiesd`` module
+# into ``sys.modules`` (same pattern as ``test_paged_attention`` /
+# ``benchmarks/conftest``). No production code is exercised for real kernels.
+
+
+def _npu_impl(causal: bool = False) -> FlashAttentionImpl:
+    return FlashAttentionImpl(num_heads=2, head_size=4, softmax_scale=0.5, causal=causal)
+
+
+def _fake_mindiesd(monkeypatch, *, attention_forward=None, attention_forward_varlen=None):
+    """Install a stub ``mindiesd`` so local ``from mindiesd import ...`` works.
+
+    Restored automatically by monkeypatch. Defaults are no-op callables so the
+    top-level import in ``forward_fa_npu`` succeeds even when the test mocks the
+    instance methods that would otherwise call them.
+    """
+    fake = SimpleNamespace(
+        attention_forward=attention_forward or Mock(return_value=torch.zeros(1)),
+        attention_forward_varlen=attention_forward_varlen or Mock(return_value=torch.zeros(1)),
+    )
+    monkeypatch.setitem(sys.modules, "mindiesd", fake)
+    return fake
+
+
+# --- Test group A: boundary resolution (_resolve_packed_seq_npu) -------------
+
+
+def test_resolve_packed_seq_accepts_real_plus_pad_contract():
+    impl = _npu_impl()
+    q = torch.randn(1, 8, 2, 4)
+    cu = torch.tensor([0, 5, 8], dtype=torch.int32)
+    extra = {"cu_seqlens_q": cu, "cu_seqlens_k": cu, "max_seqlen_q": 5, "max_seqlen_k": 5}
+
+    assert impl._resolve_packed_seq_npu(q, q, extra) == ([5, 8], [5, 8])
+
+
+def test_resolve_packed_seq_single_document_no_padding():
+    impl = _npu_impl()
+    q = torch.randn(1, 8, 2, 4)
+    cu = torch.tensor([0, 8], dtype=torch.int32)
+    extra = {"cu_seqlens_q": cu, "cu_seqlens_k": cu, "max_seqlen_q": 8, "max_seqlen_k": 8}
+
+    assert impl._resolve_packed_seq_npu(q, q, extra) == ([8], [8])
+
+
+@pytest.mark.parametrize(
+    "case",
+    [
+        "causal",
+        "missing_key",
+        "batch_ne_1",
+        "cu_too_long",
+        "cu_qk_mismatch",
+        "max_seqlen_not_int",
+        "used_gt_total",
+        "used_zero",
+        "pad_longer_than_real",
+    ],
+)
+def test_resolve_packed_seq_rejects_malformed_metadata(case):
+    """Any metadata outside the exact [real, pad] contract returns None so the
+    caller falls back to the masked path."""
+    cu3 = torch.tensor([0, 5, 8], dtype=torch.int32)
+    cu2 = torch.tensor([0, 8], dtype=torch.int32)
+    cu4 = torch.tensor([0, 3, 5, 8], dtype=torch.int32)
+    cu_short = torch.tensor([0, 3, 8], dtype=torch.int32)
+    base = {"cu_seqlens_q": cu3, "cu_seqlens_k": cu3, "max_seqlen_q": 5, "max_seqlen_k": 5}
+    q = torch.randn(1, 8, 2, 4)
+
+    impl = _npu_impl(causal=(case == "causal"))
+    if case == "missing_key":
+        extra = {"cu_seqlens_q": cu3, "max_seqlen_q": 5}  # 3 of 4 keys absent
+    elif case == "batch_ne_1":
+        q = torch.randn(2, 8, 2, 4)
+        extra = base
+    elif case == "cu_too_long":
+        extra = {**base, "cu_seqlens_q": cu4, "cu_seqlens_k": cu4}
+    elif case == "cu_qk_mismatch":
+        extra = {**base, "cu_seqlens_q": cu3, "cu_seqlens_k": cu2}
+    elif case == "max_seqlen_not_int":
+        extra = {**base, "max_seqlen_q": 5.0}
+    elif case == "used_gt_total":
+        extra = {**base, "max_seqlen_q": 9}  # 9 > total length 8
+    elif case == "used_zero":
+        extra = {**base, "max_seqlen_q": 0, "max_seqlen_k": 0}
+    elif case == "pad_longer_than_real":
+        # real=3, pad=5: the real document must be the longer one.
+        extra = {
+            "cu_seqlens_q": cu_short,
+            "cu_seqlens_k": cu_short,
+            "max_seqlen_q": 3,
+            "max_seqlen_k": 3,
+        }
+    else:
+        extra = base
+
+    assert impl._resolve_packed_seq_npu(q, q, extra) is None
+
+
+# --- Test group B: env dispatch in forward_fa_npu (current behavior) --------
+
+
+def _npu_impl_with_mocked_paths() -> tuple[FlashAttentionImpl, torch.Tensor]:
+    """FlashAttentionImpl whose two packed paths are Mocks returning a sentinel.
+
+    Lets a test assert which branch ``forward_fa_npu`` dispatched to without
+    running any real kernel.
+    """
+    impl = _npu_impl()
+    sentinel = torch.tensor([1.0])
+    impl._forward_varlen_packed_npu = Mock(return_value=sentinel)
+    impl._forward_prefix_kv_slice_npu = Mock(return_value=sentinel)
+    return impl, sentinel
+
+
+def test_npu_env_unset_uses_varlen(monkeypatch):
+    monkeypatch.delenv("MINDIE_SD_FA_TYPE", raising=False)
+    _fake_mindiesd(monkeypatch)
+    impl, _ = _npu_impl_with_mocked_paths()
+    metadata = AttentionMetadata(extra={"npu_attn_varlen": True})
+
+    impl.forward_fa_npu(torch.randn(1, 8, 2, 4), *[torch.randn(1, 8, 2, 4)] * 2, metadata)
+
+    impl._forward_varlen_packed_npu.assert_called_once()
+    impl._forward_prefix_kv_slice_npu.assert_not_called()
+
+
+def test_npu_env_laser_uses_slice(monkeypatch):
+    monkeypatch.setenv("MINDIE_SD_FA_TYPE", "ascend_laser_attention")
+    _fake_mindiesd(monkeypatch)
+    impl, _ = _npu_impl_with_mocked_paths()
+    metadata = AttentionMetadata(extra={"npu_attn_varlen": True})
+
+    impl.forward_fa_npu(torch.randn(1, 8, 2, 4), *[torch.randn(1, 8, 2, 4)] * 2, metadata)
+
+    impl._forward_prefix_kv_slice_npu.assert_called_once()
+    impl._forward_varlen_packed_npu.assert_not_called()
+
+
+@pytest.mark.parametrize("fa_type", ["prompt_flash_attn", "fused_attn_score", "ascend_laser_attenton", "garbage"])
+def test_npu_env_non_laser_value_falls_back_to_varlen(monkeypatch, fa_type):
+    """Locks in the CURRENT dispatch behavior (PR #5891).
+
+    Any value other than ``ascend_laser_attention`` — including the valid
+    ``prompt_flash_attn``/``fused_attn_score`` and typos — is silently routed to
+    the TND varlen op, which cannot honor ``MINDIE_SD_FA_TYPE``. This is review
+    comment #1 on the PR and is intentionally captured here as a regression
+    net: update this test when the dispatch starts honoring/rejecting these.
+    """
+    monkeypatch.setenv("MINDIE_SD_FA_TYPE", fa_type)
+    _fake_mindiesd(monkeypatch)
+    impl, _ = _npu_impl_with_mocked_paths()
+    metadata = AttentionMetadata(extra={"npu_attn_varlen": True})
+
+    impl.forward_fa_npu(torch.randn(1, 8, 2, 4), *[torch.randn(1, 8, 2, 4)] * 2, metadata)
+
+    impl._forward_varlen_packed_npu.assert_called_once()
+    impl._forward_prefix_kv_slice_npu.assert_not_called()
+
+
+def test_npu_varlen_opt_in_unset_takes_mask_path(monkeypatch):
+    """Models that do not set ``npu_attn_varlen`` (e.g. Wan/Cosmos) keep using
+    the existing masked attention_forward path; the new branches are skipped."""
+    fake_forward = Mock(return_value=torch.zeros(1, 8, 2, 4))
+    _fake_mindiesd(monkeypatch, attention_forward=fake_forward)
+    impl, _ = _npu_impl_with_mocked_paths()
+    metadata = AttentionMetadata(
+        attn_mask=torch.tensor([[True, True, True, True, True, False, False, False]]),
+        extra={},  # no npu_attn_varlen opt-in
+    )
+    q = torch.randn(1, 8, 2, 4)
+
+    out = impl.forward_fa_npu(q, q, q, metadata)
+
+    impl._forward_varlen_packed_npu.assert_not_called()
+    impl._forward_prefix_kv_slice_npu.assert_not_called()
+    fake_forward.assert_called_once()
+    assert out is fake_forward.return_value
+
+
+# --- Test group C: laser input pre-scaling in _forward_prefix_kv_slice_npu --
+
+
+def test_prefix_kv_slice_applies_laser_input_scaling(monkeypatch):
+    monkeypatch.setenv("MINDIE_SD_FA_TYPE", "ascend_laser_attention")
+    captured: dict = {}
+
+    def fake_attention_forward(query, key, value, **kwargs):
+        captured["query"] = query
+        captured["key"] = key
+        captured["value"] = value
+        captured.update(kwargs)
+        return torch.ones_like(query)  # deterministic stand-in kernel output
+
+    _fake_mindiesd(monkeypatch, attention_forward=fake_attention_forward)
+    impl = _npu_impl()  # softmax_scale == 0.5
+    q = torch.randn(1, 8, 2, 4)
+    cu = torch.tensor([0, 5, 8], dtype=torch.int32)
+    extra = {
+        "cu_seqlens_q": cu,
+        "cu_seqlens_k": cu,
+        "max_seqlen_q": 5,
+        "max_seqlen_k": 5,
+        "laser_input_scale": 256.0,
+    }
+
+    out = impl._forward_prefix_kv_slice_npu(q, q, q, extra)
+
+    # K/V sliced to the valid prefix (used_k == 5); q keeps full length.
+    assert captured["key"].shape == (1, 5, 2, 4)
+    assert captured["value"].shape == (1, 5, 2, 4)
+    assert captured["query"].shape == (1, 8, 2, 4)
+    assert captured["attn_mask"] is None
+    # Pre-divide q/k/v by the factor and compensate the kernel scale by its
+    # square (exact for power-of-two); output scaled back by the factor.
+    torch.testing.assert_close(captured["query"], q / 256.0)
+    torch.testing.assert_close(captured["key"], q[:, :5] / 256.0)
+    assert captured["scale"] == pytest.approx(0.5 * 256.0**2)
+    torch.testing.assert_close(out, torch.ones(1, 8, 2, 4) * 256.0)
+
+
+def test_prefix_kv_slice_no_scaling_without_factor(monkeypatch):
+    monkeypatch.setenv("MINDIE_SD_FA_TYPE", "ascend_laser_attention")
+    captured: dict = {}
+
+    def fake_attention_forward(query, key, value, **kwargs):
+        captured["query"] = query
+        captured.update(kwargs)
+        return torch.ones_like(query)
+
+    _fake_mindiesd(monkeypatch, attention_forward=fake_attention_forward)
+    impl = _npu_impl()
+    q = torch.randn(1, 8, 2, 4)
+    cu = torch.tensor([0, 5, 8], dtype=torch.int32)
+    extra = {
+        "cu_seqlens_q": cu,
+        "cu_seqlens_k": cu,
+        "max_seqlen_q": 5,
+        "max_seqlen_k": 5,
+        # no laser_input_scale → no pre-scaling
+    }
+
+    out = impl._forward_prefix_kv_slice_npu(q, q, q, extra)
+
+    torch.testing.assert_close(captured["query"], q)
+    assert captured["scale"] == pytest.approx(0.5)
+    torch.testing.assert_close(out, torch.ones(1, 8, 2, 4))
+
+
 if __name__ == "__main__":
    print("Running FlashAttention Padding Tests...")
    print("=" * 60)
@@ -5,10 +5,9 @@

 The grids here are the ones MiniMax-H3 FL2VA actually produces for an 8.7s clip:
 1280x768 gives a 62x24x40 latent grid whose 59520 video rows land on the 128-row
-kernel block, and 1344x768 gives 62x24x42 whose 62496 rows straddle it. Only the
-first can run sparse: rf_v2 tiles the sequence as a whole but pools its block
-mask from the video and prefix segments separately, so the two only agree on the
-block count when the video segment ends on a boundary.
+kernel block, and 1344x768 gives 62x24x42 whose 62496 rows do not. Both are
+handed to MindIE-SD: its rf_v2 preprocessing promotes an irregular video tail
+to the always-kept prefix segment before generating the block mask.
 """

 import dataclasses
@@ -57,10 +56,13 @@ def test_block_aligned_video_segment_runs_sparse():
    assert plan.latent_shape == list(ALIGNED_GRID)


-def test_misaligned_video_segment_stays_dense():
-    # Realigning the sequence would mean padding it, and rf_v2 ignores attn_mask:
-    # pad keys would take a share of every real query's softmax denominator.
-    assert make_impl()._resolve_plan(make_metadata(MISALIGNED_GRID)) is None
+def test_misaligned_video_segment_runs_sparse_with_updated_mindiesd():
+    plan = make_impl()._resolve_plan(make_metadata(MISALIGNED_GRID))
+
+    assert plan is not None
+    assert plan.prefix_len == PREFIX_ROWS
+    assert plan.used_len == PREFIX_ROWS + 62496
+    assert plan.latent_shape == list(MISALIGNED_GRID)


@pytest.mark.parametrize("prefix_len", [1, 127, _BLOCK_SIZE, 710, 900])
@@ -155,3 +157,28 @@ def test_fully_populated_mask_reproduces_dense_attention():
    ).transpose(1, 2)
    error = (out.float() - reference.float()).abs().mean() / reference.float().abs().mean()
    assert error < 2e-3, f"mean relative error {error:.4%} against dense attention"
+
+
+@pytest.mark.skipif(not current_omni_platform.is_npu(), reason="rf_v2 runs on Ascend NPU only.")
+def test_irregular_video_tail_reproduces_dense_attention_with_updated_mindiesd():
+    """vLLM forwards the full irregular grid; MindIE-SD owns its tail handling."""
+    torch.manual_seed(1)
+    prefix_len, grid = 384, (4, 16, 10)
+    video_len = grid[0] * grid[1] * grid[2]
+    used = prefix_len + video_len
+    heads, head_dim = 4, 128
+    q, k, v = (torch.randn(1, used, heads, head_dim, dtype=torch.bfloat16, device="npu") for _ in range(3))
+
+    impl = make_impl()
+    impl.rainfusion = dataclasses.replace(impl.rainfusion, sparsity=0.0)
+    plan = RainFusionPlan(prefix_len=prefix_len, used_len=used, latent_shape=list(grid))
+    out = impl._forward_sparse_npu(q, k, v, plan)
+
+    reference = torch.nn.functional.scaled_dot_product_attention(
+        q.transpose(1, 2),
+        k.transpose(1, 2),
+        v.transpose(1, 2),
+        scale=impl.softmax_scale,
+    ).transpose(1, 2)
+    error = (out.float() - reference.float()).abs().mean() / reference.float().abs().mean()
+    assert error < 2e-3, f"mean relative error {error:.4%} against dense attention"
@@ -471,6 +471,66 @@ class TestTeaCacheBackend:
        assert backend.enabled is True
        mock_apply_hook.assert_called_once()

+    @patch("vllm_omni.diffusion.cache.teacache.backend.apply_teacache_hook")
+    def test_enable_uses_generic_default_threshold(self, mock_apply_hook):
+        pipeline = Mock()
+        pipeline.__class__.__name__ = "QwenImagePipeline"
+        pipeline.transformer = Mock()
+        pipeline.transformer.__class__.__name__ = "QwenImageTransformer2DModel"
+
+        TeaCacheBackend(DiffusionCacheConfig()).enable(pipeline)
+        assert mock_apply_hook.call_args.args[1].rel_l1_thresh == 0.2
+
+    @pytest.mark.parametrize("partition", ["fl2va", "combined"])
+    @pytest.mark.parametrize(
+        ("configured_threshold", "expected_threshold"),
+        [(None, 0.17), (0.2, 0.2)],
+    )
+    @patch("vllm_omni.diffusion.cache.teacache.backend.apply_teacache_hook")
+    def test_minimax_h3_only_enables_fl2va_teacache(
+        self,
+        mock_apply_hook,
+        partition,
+        configured_threshold,
+        expected_threshold,
+    ):
+        pipeline = Mock()
+        pipeline.__class__.__name__ = "MiniMaxH3Pipeline"
+        pipeline.partition = partition
+        pipeline.transformer = Mock()
+        pipeline.transformer.__class__.__name__ = "MiniMaxH3DiTModel"
+        pipeline.transformers_ref = Mock()
+
+        config = (
+            DiffusionCacheConfig()
+            if configured_threshold is None
+            else DiffusionCacheConfig(rel_l1_thresh=configured_threshold)
+        )
+        backend = TeaCacheBackend(config)
+        backend.enable(pipeline)
+
+        mock_apply_hook.assert_called_once()
+        transformer, teacache_config = mock_apply_hook.call_args.args
+        assert transformer is pipeline.transformer
+        assert transformer is not pipeline.transformers_ref
+        assert teacache_config.rel_l1_thresh == expected_threshold
+
+    @patch("vllm_omni.diffusion.cache.teacache.backend.apply_teacache_hook")
+    def test_minimax_h3_rejects_ref2va_teacache(self, mock_apply_hook):
+        pipeline = Mock()
+        pipeline.__class__.__name__ = "MiniMaxH3Pipeline"
+        pipeline.partition = "ref2va"
+        pipeline.transformer = Mock()
+        pipeline.transformer.__class__.__name__ = "MiniMaxH3DiTModel"
+
+        backend = TeaCacheBackend(DiffusionCacheConfig())
+
+        with pytest.raises(ValueError, match="only supports the MiniMax-H3 FL2VA partition"):
+            backend.enable(pipeline)
+
+        assert backend.enabled is False
+        mock_apply_hook.assert_not_called()
+
    @patch("vllm_omni.diffusion.cache.teacache.backend.apply_teacache_hook")
    def test_enable_with_coefficients(self, mock_apply_hook):
        """Test enabling TeaCache with custom coefficients."""
@@ -15,19 +15,23 @@ Currently implemented:
 - TestFlux2KleinExtractor: Flux2Klein model extractor
 - TestFlux2Extractor: Flux2 model extractor
 - TestFluxExtractor: Flux model extractor
+- TestMiniMaxH3Extractor: MiniMaxH3DiTModel extractor
 """

 from abc import ABC, abstractmethod
+from types import SimpleNamespace
 from unittest.mock import MagicMock, Mock, patch

 import pytest
 import torch
+import torch.nn as nn

 from tests.helpers.mark import hardware_test
 from vllm_omni.diffusion.cache.teacache.extractors import (
    extract_flux2_context,
    extract_flux2_klein_context,
    extract_flux_context,
+    extract_minimax_h3_context,
 )
 from vllm_omni.diffusion.models.flux.flux_transformer import FluxTransformer2DModel
 from vllm_omni.diffusion.models.flux2_klein.flux2_klein_transformer import (
@@ -423,3 +427,288 @@ class TestFluxExtractor(BaseExtractorTest):
                img_ids=torch.randint(0, 64, (1, 16, 3)),
                txt_ids=torch.randint(0, 64, (1, 8, 3)),
            )
+
+
+class _MiniMaxH3FakeLinear(nn.Module):
+    def __init__(
+        self,
+        in_features,
+        out_features,
+        *,
+        bias=False,
+        params_dtype=None,
+        quant_config=None,
+        prefix="",
+        total_num_heads=None,
+        total_num_kv_heads=None,
+        **kwargs,
+    ):
+        del kwargs
+        super().__init__()
+        self.prefix = prefix
+        self.quant_config = quant_config
+        self.weight = nn.Parameter(torch.randn(out_features, in_features, dtype=params_dtype or torch.float32))
+        if bias:
+            self.bias = nn.Parameter(torch.zeros(out_features, dtype=params_dtype or torch.float32))
+        else:
+            self.register_parameter("bias", None)
+        self.num_heads = total_num_heads
+        self.num_kv_heads = total_num_kv_heads
+
+    def forward(self, x):
+        return torch.nn.functional.linear(x, self.weight, self.bias), None
+
+
+class _MiniMaxH3FakeMergedLinear(_MiniMaxH3FakeLinear):
+    def __init__(self, in_features, out_features, **kwargs):
+        super().__init__(in_features, sum(out_features), **kwargs)
+
+
+class _MiniMaxH3FakeQKVLinear(_MiniMaxH3FakeLinear):
+    def __init__(
+        self,
+        *,
+        hidden_size,
+        head_size,
+        total_num_heads,
+        total_num_kv_heads,
+        **kwargs,
+    ):
+        qkv_features = (total_num_heads + 2 * total_num_kv_heads) * head_size
+        super().__init__(
+            hidden_size,
+            qkv_features,
+            total_num_heads=total_num_heads,
+            total_num_kv_heads=total_num_kv_heads,
+            **kwargs,
+        )
+
+
+class _MiniMaxH3FakeAttention(nn.Module):
+    class FakeBackend:
+        supports_prefix_kv_slicing = False
+
+        # Mirror AttentionBackend.supports_packed_mask_free: H3's packed
+        # attention path calls attn_backend.supports_packed_mask_free() to decide
+        # whether to build the padding mask. Return False so the fake exercises
+        # the masked path (the abstract backend default).
+        @classmethod
+        def supports_packed_mask_free(cls) -> bool:
+            return False
+
+    def __init__(self, **kwargs):
+        del kwargs
+        super().__init__()
+        self.attn_backend = self.FakeBackend
+        self.use_ring = False
+
+    def forward(self, q, k, v, metadata=None):
+        del k, v, metadata
+        return q
+
+
+def _minimax_h3_small_od_config():
+    arch = {
+        "num_layers": 1,
+        "token_refiner_num_layers": 1,
+        "hidden_size": 8,
+        "num_attention_heads": 1,
+        "attention_head_dim": 8,
+        "ffn_hidden_size": 16,
+        "latents_dim": 2,
+        "audio_latents_dim": 2,
+        "patch_size": (1, 2, 2),
+        "text_dim": 6,
+        "timestep_input_dim": 4,
+        "time_embed_hidden_size": 8,
+        "time_embed_dim": 4,
+        "adaln_out_features": 18 * 8,
+        "final_adaln_out_features": 2 * 8,
+        "rope_inv_freq_len": 1,
+    }
+    return SimpleNamespace(
+        tf_model_config=arch,
+        parallel_config=SimpleNamespace(ulysses_degree=1),
+    )
+
+
+def _minimax_h3_sample_kwargs(*, seq_len: int = 8, device: torch.device | str = "cpu") -> dict:
+    video_row_dim = 2 * 1 * 2 * 2
+    audio_row_dim = 2
+    text_len = 2
+    img_rows = 2
+
+    img_pos = torch.tensor([2, 3], dtype=torch.long, device=device)
+    audio_pos = torch.tensor([4, 5], dtype=torch.long, device=device)
+    text_pos = torch.tensor([0, 1], dtype=torch.long, device=device)
+    infer_out_pos = img_pos.clone()
+
+    token_tags = torch.full((seq_len,), -1, dtype=torch.long, device=device)
+    token_tags[text_pos] = 1
+    token_tags[img_pos] = 0
+    token_tags[audio_pos] = 2
+
+    inverse_indices = torch.zeros(seq_len, dtype=torch.long, device=device)
+
+    return {
+        "x": torch.randn(1, seq_len, video_row_dim, device=device),
+        "audio_x": torch.randn(1, seq_len, audio_row_dim, device=device),
+        "img_position_ids": torch.zeros(1, seq_len, 3, dtype=torch.float64, device=device),
+        "unique_timesteps": torch.tensor([0.5], dtype=torch.float32, device=device),
+        "inverse_indices": inverse_indices,
+        "update_mask": torch.ones(img_rows, dtype=torch.bool, device=device),
+        "token_tags": token_tags,
+        "prompt_embeds": torch.randn(text_len, 6, device=device),
+        "img_pos_info": {"position_ids": img_pos},
+        "audio_pos_info": {"position_ids": audio_pos},
+        "text_pos_info": {"position_ids": text_pos},
+        "img_pos_for_infer_output_info": {"position_ids": infer_out_pos},
+        "packed_seq_params": {
+            "cu_seqlens_q": torch.tensor([0, seq_len], dtype=torch.int32, device=device),
+            "max_seqlen_q": seq_len,
+        },
+        "refiner_packed_seq_params": {
+            "cu_seqlens_q": torch.tensor([0, text_len], dtype=torch.int32, device=device),
+            "max_seqlen_q": text_len,
+        },
+    }
+
+
+@pytest.mark.cpu
+class TestMiniMaxH3Extractor(BaseExtractorTest):
+    """Test extract_minimax_h3_context function."""
+
+    def get_extractor(self):
+        return extract_minimax_h3_context
+
+    @pytest.fixture
+    def minimax_h3_module(self, monkeypatch):
+        """Create a minimal MiniMaxH3DiTModel with fake parallel linears for CPU tests."""
+        from vllm_omni.diffusion.models.minimax_h3 import minimax_h3_transformer as h3
+
+        monkeypatch.setattr(h3, "ColumnParallelLinear", _MiniMaxH3FakeLinear)
+        monkeypatch.setattr(h3, "MergedColumnParallelLinear", _MiniMaxH3FakeMergedLinear)
+        monkeypatch.setattr(h3, "QKVParallelLinear", _MiniMaxH3FakeQKVLinear)
+        monkeypatch.setattr(h3, "RowParallelLinear", _MiniMaxH3FakeLinear)
+        monkeypatch.setattr(h3, "Attention", _MiniMaxH3FakeAttention)
+        monkeypatch.setattr(h3, "get_tensor_model_parallel_world_size", lambda: 1)
+
+        model = h3.MiniMaxH3DiTModel(_minimax_h3_small_od_config(), quant_config=None)
+        for submodule in model.modules():
+            if isinstance(submodule, h3.MiniMaxH3Attention):
+                submodule.rope._forward_method = submodule.rope.forward_native
+        model.eval()
+        return model
+
+    def get_module(self, minimax_h3_module):
+        return minimax_h3_module
+
+    @pytest.fixture
+    def sample_inputs(self):
+        """Create sample packed forward kwargs for MiniMaxH3DiTModel."""
+        return _minimax_h3_sample_kwargs(seq_len=8)
+
+    def get_sample_inputs(self, sample_inputs):
+        return sample_inputs
+
+    def test_modulated_input_shape(self, minimax_h3_module, sample_inputs):
+        """Test that modulated_input matches packed sequence shape [T, H]."""
+        context = extract_minimax_h3_context(minimax_h3_module, **sample_inputs)
+
+        seq_len = sample_inputs["x"].shape[1]
+        assert context.modulated_input.shape == (seq_len, minimax_h3_module.hidden_size)
+
+    def test_run_transformer_blocks_callable(self, minimax_h3_module, sample_inputs):
+        """Test that run_transformer_blocks is callable."""
+        context = extract_minimax_h3_context(minimax_h3_module, **sample_inputs)
+        assert callable(context.run_transformer_blocks)
+
+    def test_postprocess_callable(self, minimax_h3_module, sample_inputs):
+        """Test that postprocess is callable."""
+        context = extract_minimax_h3_context(minimax_h3_module, **sample_inputs)
+        assert callable(context.postprocess)
+
+    def test_postprocess_output_shape(self, minimax_h3_module, sample_inputs):
+        """Test that postprocess returns video/audio logits for infer rows."""
+        context = extract_minimax_h3_context(minimax_h3_module, **sample_inputs)
+        (hidden,) = context.run_transformer_blocks()
+        video_logits, audio_logits = context.postprocess(hidden)
+
+        num_video_rows = sample_inputs["img_pos_info"]["position_ids"].shape[0]
+        num_audio_rows = sample_inputs["audio_pos_info"]["position_ids"].shape[0]
+        assert video_logits.shape[0] == num_video_rows
+        assert audio_logits.shape[0] == num_audio_rows
+
+    def test_extractor_matches_native_forward(self, minimax_h3_module, sample_inputs):
+        """The uncached extractor path must preserve the model's forward result."""
+        expected_video, expected_audio = minimax_h3_module(**sample_inputs)
+
+        context = extract_minimax_h3_context(minimax_h3_module, **sample_inputs)
+        (hidden,) = context.run_transformer_blocks()
+        actual_video, actual_audio = context.postprocess(hidden)
+
+        torch.testing.assert_close(actual_video, expected_video)
+        torch.testing.assert_close(actual_audio, expected_audio)
+
+    def test_run_transformer_blocks_forwards_packed_layout(
+        self,
+        minimax_h3_module,
+        sample_inputs,
+        monkeypatch,
+    ):
+        """The extractor block call must mirror the native packed contract."""
+        from vllm_omni.diffusion.attention.backends.abstract import VideoTokenLayout
+
+        video_layout = VideoTokenLayout(prefix_len=2, latent_grid=(1, 1, 2))
+        inputs = {**sample_inputs, "video_token_layout": video_layout}
+        block = minimax_h3_module.blocks[0]
+        original_forward = block.forward
+        received: list[tuple[int, VideoTokenLayout | None]] = []
+
+        def capture_forward(*args, **kwargs):
+            received.append((kwargs["packed_total"], kwargs["video_layout"]))
+            return original_forward(*args, **kwargs)
+
+        monkeypatch.setattr(block, "forward", capture_forward)
+        context = extract_minimax_h3_context(minimax_h3_module, **inputs)
+        context.run_transformer_blocks()
+
+        assert received == [(sample_inputs["x"].shape[1], video_layout)]
+
+    def test_teacache_reuses_h3_block_residual(self, minimax_h3_module, sample_inputs, monkeypatch):
+        """A guaranteed cache hit must skip the H3 transformer block."""
+        from vllm_omni.diffusion.cache.teacache.config import TeaCacheConfig
+        from vllm_omni.diffusion.cache.teacache.hook import TeaCacheHook
+
+        block = minimax_h3_module.blocks[0]
+        original_forward = block.forward
+        block_calls = 0
+
+        def counted_forward(*args, **kwargs):
+            nonlocal block_calls
+            block_calls += 1
+            return original_forward(*args, **kwargs)
+
+        monkeypatch.setattr(block, "forward", counted_forward)
+        hook = TeaCacheHook(
+            TeaCacheConfig(
+                transformer_type="MiniMaxH3DiTModel",
+                rel_l1_thresh=0.3,
+            )
+        )
+        hook.initialize_hook(minimax_h3_module)
+
+        hook.new_forward(minimax_h3_module, **sample_inputs)
+        video_logits, audio_logits = hook.new_forward(minimax_h3_module, **sample_inputs)
+
+        assert block_calls == 1
+        assert video_logits.shape[0] == sample_inputs["img_pos_info"]["position_ids"].shape[0]
+        assert audio_logits.shape[0] == sample_inputs["audio_pos_info"]["position_ids"].shape[0]
+
+    def test_invalid_module_raises_error(self):
+        """Test that invalid module without blocks raises ValueError."""
+        invalid_module = Mock()
+        invalid_module.blocks = []
+
+        with pytest.raises(ValueError, match="Module must have blocks"):
+            extract_minimax_h3_context(invalid_module, **_minimax_h3_sample_kwargs())
@@ -0,0 +1,52 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+import pytest
+
+from vllm_omni.diffusion.data import OmniDiffusionConfig
+from vllm_omni.diffusion.diffusion_kv.config import DiffusionKVCacheMode
+
+pytestmark = [pytest.mark.core_model, pytest.mark.cpu, pytest.mark.diffusion]
+
+
+@pytest.fixture(autouse=True)
+def _fixed_master_port(monkeypatch) -> None:
+    monkeypatch.setattr(OmniDiffusionConfig, "_resolve_master_port", lambda _self: 29500)
+
+
+def test_dense_legacy_is_default() -> None:
+    config = OmniDiffusionConfig.from_kwargs()
+
+    assert config.diffusion_kv_mode is DiffusionKVCacheMode.DENSE_LEGACY
+
+
+def test_paged_worker_local_is_rejected_until_implemented() -> None:
+    with pytest.raises(ValueError, match="reserved but not implemented"):
+        OmniDiffusionConfig.from_kwargs(diffusion_kv_mode="paged_worker_local")
+
+
+def test_unknown_cache_mode_is_rejected() -> None:
+    with pytest.raises(ValueError, match="Unsupported Diffusion KV diffusion_kv_mode"):
+        OmniDiffusionConfig.from_kwargs(diffusion_kv_mode="unknown")
+
+
+def test_non_mapping_omni_kv_config_is_rejected() -> None:
+    with pytest.raises(TypeError, match="omni_kv_config must be a mapping"):
+        OmniDiffusionConfig.from_kwargs(omni_kv_config="paged_scheduler")
+
+
+def test_paged_scheduler_rejects_dense_legacy_kv_receive() -> None:
+    with pytest.raises(ValueError, match="does not support imported AR KV"):
+        OmniDiffusionConfig.from_kwargs(
+            diffusion_kv_mode="paged_scheduler",
+            omni_kv_config={"need_recv_cache": True},
+        )
+
+
+def test_paged_scheduler_does_not_depend_on_model_registry() -> None:
+    config = OmniDiffusionConfig.from_kwargs(
+        model_class_name="FutureDiffusionModel",
+        diffusion_kv_mode="paged_scheduler",
+    )
+
+    assert config.diffusion_kv_mode is DiffusionKVCacheMode.PAGED_SCHEDULER
@@ -0,0 +1,110 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+import pytest
+
+from vllm_omni.diffusion.diffusion_kv.metadata import (
+    DiffusionKVContextMetadata,
+    DiffusionKVMetadata,
+    DiffusionKVSequenceMetadata,
+)
+
+pytestmark = [pytest.mark.diffusion, pytest.mark.core_model, pytest.mark.cpu]
+
+
+def test_diffusion_kv_metadata_uses_native_cache_group_block_ids() -> None:
+    context = DiffusionKVContextMetadata(
+        context_id="text",
+        cache_role="cross_attention",
+        num_tokens=3,
+        block_ids=([7], [11]),
+    )
+    sequence = DiffusionKVSequenceMetadata(
+        sequence_id=1,
+        prefix_len=4,
+        target_len=2,
+        seq_len=8,
+        block_ids=([1, 2], [5, 6]),
+        context_ids=(context.context_id,),
+    )
+    metadata = DiffusionKVMetadata(
+        request_id="req-0",
+        allocation_generation=3,
+        sequences=(sequence,),
+        contexts=(context,),
+    )
+
+    assert metadata.sequences[0].block_ids == ([1, 2], [5, 6])
+    assert metadata.contexts[0].block_ids == ([7], [11])
+
+
+def test_cfg_sequences_can_share_one_request_context() -> None:
+    shared_context = DiffusionKVContextMetadata(
+        context_id="shared-text",
+        cache_role="cross_attention",
+        num_tokens=3,
+        block_ids=([7], [11]),
+    )
+    sequences = tuple(
+        DiffusionKVSequenceMetadata(
+            sequence_id=sequence_id,
+            prefix_len=4,
+            target_len=2,
+            seq_len=8,
+            block_ids=([sequence_id + 1],),
+            context_ids=(shared_context.context_id,),
+        )
+        for sequence_id in range(2)
+    )
+
+    metadata = DiffusionKVMetadata(
+        request_id="req-shared-context",
+        allocation_generation=1,
+        sequences=sequences,
+        contexts=(shared_context,),
+    )
+
+    assert metadata.contexts == (shared_context,)
+    assert metadata.sequences[0].context_ids == (shared_context.context_id,)
+    assert metadata.sequences[1].context_ids == (shared_context.context_id,)
+
+
+def test_cfg_sequences_can_reference_branch_specific_contexts() -> None:
+    conditional_context = DiffusionKVContextMetadata(
+        context_id="conditional-text",
+        cache_role="cross_attention",
+        num_tokens=3,
+        block_ids=([7], [11]),
+    )
+    unconditional_context = DiffusionKVContextMetadata(
+        context_id="unconditional-text",
+        cache_role="cross_attention",
+        num_tokens=3,
+        block_ids=([8], [12]),
+    )
+    metadata = DiffusionKVMetadata(
+        request_id="req-branch-contexts",
+        allocation_generation=1,
+        sequences=(
+            DiffusionKVSequenceMetadata(
+                sequence_id=0,
+                prefix_len=4,
+                target_len=2,
+                seq_len=8,
+                block_ids=([1],),
+                context_ids=(conditional_context.context_id,),
+            ),
+            DiffusionKVSequenceMetadata(
+                sequence_id=1,
+                prefix_len=4,
+                target_len=2,
+                seq_len=8,
+                block_ids=([2],),
+                context_ids=(unconditional_context.context_id,),
+            ),
+        ),
+        contexts=(conditional_context, unconditional_context),
+    )
+
+    assert metadata.sequences[0].context_ids == (conditional_context.context_id,)
+    assert metadata.sequences[1].context_ids == (unconditional_context.context_id,)
@@ -0,0 +1,200 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+import pytest
+import torch
+import vllm.v1.core.single_type_kv_cache_manager as native_kv_managers
+from vllm.v1.core.kv_cache_manager import KVCacheManager
+from vllm.v1.core.kv_cache_utils import BlockHash
+from vllm.v1.kv_cache_interface import (
+    FullAttentionSpec,
+    KVCacheConfig,
+    KVCacheGroupSpec,
+    KVCacheTensor,
+)
+from vllm.v1.request import RequestStatus
+
+from vllm_omni.diffusion.diffusion_kv.request import DiffusionKVContext, DiffusionKVRequest
+
+pytestmark = [pytest.mark.core_model, pytest.mark.cpu, pytest.mark.diffusion]
+
+BLOCK_SIZE = 4
+
+
+def _request(**overrides) -> DiffusionKVRequest:
+    values = dict(
+        request_id="req/diffusion-kv/0",
+        sequence_id=0,
+        prefix_len=4,
+        target_len=4,
+        seq_len=8,
+    )
+    values.update(overrides)
+    return DiffusionKVRequest(**values)
+
+
+def test_request_exposes_native_request_and_diffusion_semantics() -> None:
+    request = _request()
+
+    assert request.sequence_id == 0
+    assert request.prefix_len == 4
+    assert request.target_len == 4
+    assert request.num_tokens == request.seq_len == 8
+    assert request.num_prompt_tokens == 4
+    assert request.num_computed_tokens == 0
+    assert request.block_hashes == []
+    assert request.kv_contexts == ()
+    assert request.skip_reading_prefix_cache is True
+    assert request.shared_prefix_boundary == 0
+    assert request.status is RequestStatus.WAITING
+
+
+@pytest.mark.parametrize(
+    ("overrides", "message"),
+    [
+        ({"sequence_id": -1}, "sequence_id"),
+        ({"prefix_len": -1}, "prefix_len"),
+        ({"target_len": 0}, "target_len"),
+        ({"seq_len": 0}, "seq_len"),
+        ({"prefix_len": 5, "target_len": 4, "seq_len": 8}, "must not exceed"),
+    ],
+)
+def test_request_rejects_invalid_lengths(overrides, message: str) -> None:
+    with pytest.raises(ValueError, match=message):
+        _request(**overrides)
+
+
+@pytest.mark.parametrize(
+    ("overrides", "message"),
+    [
+        ({"context_id": ""}, "context_id"),
+        ({"cache_role": ""}, "cache_role"),
+        ({"num_tokens": 0}, "num_tokens"),
+    ],
+)
+def test_context_rejects_invalid_identity_or_length(overrides, message: str) -> None:
+    values = dict(context_id="text", cache_role="cross.text", num_tokens=8)
+    values.update(overrides)
+
+    with pytest.raises(ValueError, match=message):
+        DiffusionKVContext(**values)
+
+
+def test_request_owns_independent_kv_contexts() -> None:
+    text = DiffusionKVContext(context_id="text", cache_role="cross.text", num_tokens=8)
+    image = DiffusionKVContext(context_id="image", cache_role="cross.image", num_tokens=4)
+
+    request = _request(kv_contexts=(text, image))
+
+    assert request.kv_contexts == (text, image)
+
+
+def test_request_rejects_duplicate_context_ids() -> None:
+    contexts = (
+        DiffusionKVContext(context_id="condition", cache_role="cross.text", num_tokens=8),
+        DiffusionKVContext(context_id="condition", cache_role="cross.image", num_tokens=4),
+    )
+
+    with pytest.raises(ValueError, match="unique context_id"):
+        _request(kv_contexts=contexts)
+
+
+def test_request_rejects_non_context_values() -> None:
+    with pytest.raises(TypeError, match="DiffusionKVContext"):
+        _request(kv_contexts=(object(),))
+
+
+def _manager(*, enable_caching: bool) -> KVCacheManager:
+    # vLLM 0.26 lazily registers its built-in KV specs. Another test may have
+    # already registered an out-of-tree spec, making the registry non-empty
+    # before FullAttentionSpec is registered. Initialize the native mappings
+    # explicitly; vLLM 0.25 has no registration hook and needs no setup.
+    register_all = getattr(native_kv_managers, "register_all_kvcache_specs", None)
+    if register_all is not None:
+        register_all(None)
+
+    spec = FullAttentionSpec(
+        block_size=BLOCK_SIZE,
+        num_kv_heads=2,
+        head_size=8,
+        dtype=torch.bfloat16,
+    )
+    group = KVCacheGroupSpec(layer_names=["layer0"], kv_cache_spec=spec)
+    config = KVCacheConfig(
+        num_blocks=8,
+        kv_cache_tensors=[KVCacheTensor(size=spec.page_size_bytes * 8, shared_by=["layer0"])],
+        kv_cache_groups=[group],
+    )
+    return KVCacheManager(
+        config,
+        max_model_len=64,
+        scheduler_block_size=BLOCK_SIZE,
+        hash_block_size=BLOCK_SIZE,
+        enable_caching=enable_caching,
+    )
+
+
+def test_empty_hash_request_conforms_to_native_request_local_allocation() -> None:
+    manager = _manager(enable_caching=True)
+    free_before = manager.block_pool.get_num_free_blocks()
+    request = _request()
+
+    # vLLM 0.26 adds a shared-prefix-boundary return value.
+    computed_blocks, num_computed_tokens, *_ = manager.get_computed_blocks(request)
+    assert num_computed_tokens == 0
+    blocks = manager.allocate_slots(
+        request,
+        num_new_tokens=request.num_tokens,
+        new_computed_blocks=computed_blocks,
+        delay_cache_blocks=True,
+        full_sequence_must_fit=True,
+    )
+
+    assert blocks is not None
+    assert len(manager.get_block_ids(request.request_id)[0]) == 2
+    # Empty hashes intentionally mean request-local residency only. Calling
+    # native cache_blocks() would require a canonical hash for the full prefix
+    # block and is therefore outside this request's contract.
+    manager.free(request)
+    assert manager.block_pool.get_num_free_blocks() == free_before
+
+
+def test_hashed_request_conforms_to_native_delayed_prefix_commit() -> None:
+    manager = _manager(enable_caching=True)
+    prefix_hash = BlockHash(b"prefix".ljust(32, b"\0"))
+    target_hash = BlockHash(b"target".ljust(32, b"\0"))
+    request = _request(block_hashes=[prefix_hash, target_hash])
+
+    computed_blocks, num_computed_tokens, *_ = manager.get_computed_blocks(request)
+    assert num_computed_tokens == 0
+    blocks = manager.allocate_slots(
+        request,
+        num_new_tokens=request.num_tokens,
+        new_computed_blocks=computed_blocks,
+        delay_cache_blocks=True,
+        full_sequence_must_fit=True,
+    )
+
+    assert blocks is not None
+    manager.cache_blocks(request, request.prefix_len)
+    manager.free(request)
+
+    cached_request = _request(
+        request_id="req/diffusion-kv/cached",
+        block_hashes=[prefix_hash, target_hash],
+    )
+    cached_blocks, num_computed_tokens, *_ = manager.get_computed_blocks(cached_request)
+
+    # Only the stable prefix was published. The dynamic target remains a miss.
+    assert num_computed_tokens == cached_request.prefix_len
+    assert len(cached_blocks.blocks[0]) == 1
+    cached_allocation = manager.allocate_slots(
+        cached_request,
+        num_new_tokens=cached_request.num_tokens - num_computed_tokens,
+        num_new_computed_tokens=num_computed_tokens,
+        new_computed_blocks=cached_blocks,
+        delay_cache_blocks=True,
+        full_sequence_must_fit=True,
+    )
+    assert cached_allocation is not None
+    manager.free(cached_request)
@@ -0,0 +1,290 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+from types import SimpleNamespace
+from unittest.mock import Mock
+
+import pytest
+
+from vllm_omni.diffusion.data import DiffusionOutput
+from vllm_omni.diffusion.diffusion_kv.config import DiffusionKVCacheMode
+from vllm_omni.diffusion.diffusion_kv.metadata import (
+    DiffusionKVMetadata,
+    DiffusionKVSequenceMetadata,
+)
+from vllm_omni.diffusion.executor.multiproc_executor import MultiprocDiffusionExecutor
+from vllm_omni.diffusion.sched.interface import (
+    CachedRequestData,
+    DiffusionSchedulerOutput,
+    NewRequestData,
+)
+from vllm_omni.diffusion.worker.diffusion_model_runner import DiffusionModelRunner
+from vllm_omni.diffusion.worker.diffusion_worker import DiffusionWorker, WorkerWrapperBase
+
+pytestmark = [pytest.mark.diffusion, pytest.mark.core_model, pytest.mark.cpu]
+
+
+def make_metadata(request_id: str = "req-0") -> DiffusionKVMetadata:
+    return DiffusionKVMetadata(
+        request_id=request_id,
+        allocation_generation=1,
+        sequences=(
+            DiffusionKVSequenceMetadata(
+                sequence_id=0,
+                prefix_len=4,
+                target_len=2,
+                seq_len=8,
+                block_ids=([1, 2],),
+            ),
+        ),
+    )
+
+
+def make_runner(mode: DiffusionKVCacheMode) -> DiffusionModelRunner:
+    runner = object.__new__(DiffusionModelRunner)
+    runner.od_config = SimpleNamespace(diffusion_kv_mode=mode)
+    return runner
+
+
+def make_scheduler_output(*new_reqs: NewRequestData) -> DiffusionSchedulerOutput:
+    return DiffusionSchedulerOutput(
+        step_id=0,
+        scheduled_new_reqs=list(new_reqs),
+        scheduled_cached_reqs=CachedRequestData.make_empty(),
+        finished_req_ids=set(),
+        num_running_reqs=len(new_reqs),
+        num_waiting_reqs=0,
+    )
+
+
+def make_executor() -> tuple[MultiprocDiffusionExecutor, list[tuple]]:
+    executor = object.__new__(MultiprocDiffusionExecutor)
+    executor.od_config = SimpleNamespace(
+        enable_distributed_layerwise_offload=False,
+        dlo_use_allgather=True,
+    )
+    executor._ensure_open = lambda: None
+    calls: list[tuple] = []
+
+    def collective_rpc(method, *, args, unique_reply_rank, exec_all_ranks):
+        calls.append((method, args, unique_reply_rank, exec_all_ranks))
+        return DiffusionOutput(output=None)
+
+    executor.collective_rpc = collective_rpc
+    return executor, calls
+
+
+def test_new_request_data_carries_scheduler_allocation_atomically() -> None:
+    req = SimpleNamespace(request_id="req-0")
+    state = SimpleNamespace(request_id="req-0", req=req)
+    metadata = make_metadata()
+
+    new_req = NewRequestData.from_state(state, diffusion_kv_metadata=metadata)
+
+    assert new_req.req is req
+    assert new_req.diffusion_kv_metadata is metadata
+
+
+def test_paged_request_without_metadata_reaches_request_forward() -> None:
+    runner = make_runner(DiffusionKVCacheMode.PAGED_SCHEDULER)
+    expected_output = DiffusionOutput(output=None)
+    runner._execute_request_list = Mock(
+        return_value=SimpleNamespace(
+            runner_outputs=[SimpleNamespace(result=expected_output)],
+        )
+    )
+    req = SimpleNamespace(request_id="req-0")
+
+    assert runner.execute_model(req) is expected_output
+
+    runner._execute_request_list.assert_called_once()
+
+
+def test_dense_request_rejects_metadata_before_request_forward() -> None:
+    runner = make_runner(DiffusionKVCacheMode.DENSE_LEGACY)
+    runner._execute_request_list = Mock()
+
+    with pytest.raises(ValueError, match="dense_legacy.*must not carry"):
+        runner.execute_model(
+            SimpleNamespace(request_id="req-0"),
+            diffusion_kv_metadata=make_metadata(),
+        )
+
+    runner._execute_request_list.assert_not_called()
+
+
+def test_paged_request_rejects_mismatched_metadata_identity() -> None:
+    runner = make_runner(DiffusionKVCacheMode.PAGED_SCHEDULER)
+    runner._execute_request_list = Mock()
+
+    with pytest.raises(ValueError, match="request mismatch"):
+        runner.execute_model(
+            SimpleNamespace(request_id="req-0"),
+            diffusion_kv_metadata=make_metadata("stale"),
+        )
+
+    runner._execute_request_list.assert_not_called()
+
+
+def test_request_rpc_rejects_envelope_request_identity_mismatch() -> None:
+    executor, calls = make_executor()
+    req = SimpleNamespace(request_id="actual-request-id")
+    new_req = NewRequestData(
+        request_id="envelope-id",
+        req=req,
+        diffusion_kv_metadata=make_metadata("envelope-id"),
+    )
+
+    with pytest.raises(ValueError, match="request identity mismatch"):
+        executor.execute_request(make_scheduler_output(new_req))
+
+    assert calls == []
+
+
+def test_batch_path_allows_missing_metadata_before_forward() -> None:
+    runner = make_runner(DiffusionKVCacheMode.PAGED_SCHEDULER)
+    expected_output = object()
+    runner._execute_request_list = Mock(return_value=expected_output)
+    new_req = NewRequestData(
+        request_id="req-0",
+        req=SimpleNamespace(request_id="req-0"),
+    )
+
+    assert runner.execute_model_batch(make_scheduler_output(new_req), runner.od_config) is expected_output
+
+    runner._execute_request_list.assert_called_once()
+
+
+def test_batch_path_rejects_envelope_request_identity_mismatch() -> None:
+    runner = make_runner(DiffusionKVCacheMode.PAGED_SCHEDULER)
+    runner._execute_request_list = Mock()
+    new_req = NewRequestData(
+        request_id="envelope-id",
+        req=SimpleNamespace(request_id="actual-request-id"),
+        diffusion_kv_metadata=make_metadata("envelope-id"),
+    )
+
+    with pytest.raises(ValueError, match="request identity mismatch"):
+        runner.execute_model_batch(make_scheduler_output(new_req), runner.od_config)
+
+    runner._execute_request_list.assert_not_called()
+
+
+def test_step_path_allows_missing_metadata_before_step_support_check() -> None:
+    runner = make_runner(DiffusionKVCacheMode.PAGED_SCHEDULER)
+    runner.pipeline = object()
+    runner._supports_step_mode = Mock(return_value=False)
+    new_req = NewRequestData(
+        request_id="req-0",
+        req=SimpleNamespace(request_id="req-0"),
+    )
+
+    with pytest.raises(ValueError, match="does not support step execution"):
+        runner.execute_stepwise(make_scheduler_output(new_req))
+
+    runner._supports_step_mode.assert_called_once()
+
+
+def test_step_path_rejects_envelope_request_identity_mismatch() -> None:
+    runner = make_runner(DiffusionKVCacheMode.PAGED_SCHEDULER)
+    runner.pipeline = object()
+    runner._supports_step_mode = Mock()
+    new_req = NewRequestData(
+        request_id="envelope-id",
+        req=SimpleNamespace(request_id="actual-request-id"),
+        diffusion_kv_metadata=make_metadata("envelope-id"),
+    )
+
+    with pytest.raises(ValueError, match="request identity mismatch"):
+        runner.execute_stepwise(make_scheduler_output(new_req))
+
+    runner._supports_step_mode.assert_not_called()
+
+
+def test_dense_request_rpc_keeps_legacy_positional_shape() -> None:
+    executor, calls = make_executor()
+    prepared_layout = object()
+    req = SimpleNamespace(request_id="req-0", prepared_layout=prepared_layout)
+    new_req = NewRequestData(request_id="req-0", req=req)
+
+    result = executor.execute_request(make_scheduler_output(new_req))
+
+    assert result.request_ids == ["req-0"]
+    assert calls == [
+        (
+            "execute_model",
+            (req, executor.od_config, None),
+            0,
+            True,
+        )
+    ]
+    assert calls[0][1][0].prepared_layout is prepared_layout
+
+
+def test_request_rpc_sends_attached_metadata_for_only_that_request() -> None:
+    executor, calls = make_executor()
+    req_0 = SimpleNamespace(request_id="req-0")
+    req_1 = SimpleNamespace(request_id="req-1")
+    metadata_0 = make_metadata("req-0")
+    metadata_1 = make_metadata("req-1")
+    new_reqs = (
+        NewRequestData(request_id="req-0", req=req_0, diffusion_kv_metadata=metadata_0),
+        NewRequestData(request_id="req-1", req=req_1, diffusion_kv_metadata=metadata_1),
+    )
+
+    result = executor.execute_request(make_scheduler_output(*new_reqs))
+
+    assert result.request_ids == ["req-0", "req-1"]
+    assert [call[1] for call in calls] == [
+        (req_0, executor.od_config, None, metadata_0),
+        (req_1, executor.od_config, None, metadata_1),
+    ]
+
+
+def test_worker_wrapper_only_extends_paged_request_rpc() -> None:
+    calls: list[tuple] = []
+
+    class Worker:
+        def execute_model(self, req, od_config, **kwargs):
+            calls.append((req, od_config, kwargs))
+            return DiffusionOutput(output=None)
+
+    wrapper = object.__new__(WorkerWrapperBase)
+    wrapper.worker = Worker()
+    req = SimpleNamespace(request_id="req-0")
+    od_config = SimpleNamespace()
+    metadata = make_metadata()
+
+    wrapper.execute_model(req, od_config)
+    wrapper.execute_model(req, od_config, diffusion_kv_metadata=metadata)
+
+    assert calls == [
+        (req, od_config, {"kv_prefetch_job": None}),
+        (
+            req,
+            od_config,
+            {
+                "kv_prefetch_job": None,
+                "diffusion_kv_metadata": metadata,
+            },
+        ),
+    ]
+
+
+def test_dense_worker_call_does_not_extend_model_runner_signature() -> None:
+    calls: list[tuple] = []
+
+    class DenseModelRunner:
+        def execute_model(self, req, kv_prefetch_job=None):
+            calls.append((req, kv_prefetch_job))
+            return DiffusionOutput(output=None)
+
+    worker = object.__new__(DiffusionWorker)
+    worker.model_runner = DenseModelRunner()
+    worker.lora_manager = None
+    worker._get_profiler = lambda: None
+    req = SimpleNamespace(request_id="req-0")
+
+    worker.execute_model(req, SimpleNamespace())
+
+    assert calls == [(req, None)]
@@ -64,6 +64,7 @@ def test_moe_ep_maps_diffusion_sp_cfg_dp_to_vllm_groups(monkeypatch):
        local_rank,
        backend,
        group_name,
+        use_all2all=False,
    ):
        del backend
        group = _FakeGroup(
@@ -73,6 +74,7 @@ def test_moe_ep_maps_diffusion_sp_cfg_dp_to_vllm_groups(monkeypatch):
            device_communicator=object(),
            reduce_scatter=lambda tensor, **kwargs: tensor,
        )
+        group.use_all2all = use_all2all
        created_groups.append(group)
        return group

@@ -121,6 +123,9 @@ def test_moe_ep_maps_diffusion_sp_cfg_dp_to_vllm_groups(monkeypatch):
    assert parallel_state.vllm_parallel_state._PCP.device_communicator is not None
    assert parallel_state.vllm_parallel_state._DP.device_communicator is not None
    assert parallel_state.vllm_parallel_state._EP.device_communicator is not None
+    # vLLM 0.27's MoE oracle asserts the EP communicator has an all2all
+    # manager, which vLLM only builds for groups created with use_all2all.
+    assert parallel_state.vllm_parallel_state._EP.use_all2all is True
    assert hasattr(parallel_state.vllm_parallel_state._PCP, "reduce_scatter")
    assert hasattr(parallel_state.vllm_parallel_state._DP, "reduce_scatter")
    assert hasattr(parallel_state.vllm_parallel_state._EP, "reduce_scatter")
@@ -313,3 +318,35 @@ def test_non_moe_ep_fails_before_vllm_ep_remap(monkeypatch):
    assert parallel_state.vllm_parallel_state._PCP is None
    assert parallel_state.vllm_parallel_state._EP is None
    assert parallel_state._EXPERT_PARALLEL_GROUP_RANKS is None
+
+
+@pytest.mark.cpu
+@pytest.mark.core_model
+def test_init_vllm_group_forwards_use_all2all_only_when_supported(monkeypatch):
+    """use_all2all reaches vLLM only on versions whose API accepts it.
+
+    vLLM 0.27 builds the EP all2all manager only for groups created with
+    use_all2all=True (its MoE oracle then asserts the manager exists —
+    build 2953, HunyuanImage3-DIT accuracy). Older vLLM has no such kwarg,
+    so forwarding it unconditionally would TypeError at init.
+    """
+    received = {}
+
+    def modern(group_ranks, local_rank, backend, group_name=None, use_device_communicator=True, use_all2all=False):
+        received["modern"] = use_all2all
+        return object()
+
+    monkeypatch.setattr(parallel_state.vllm_parallel_state, "init_model_parallel_group", modern)
+    parallel_state.init_vllm_model_parallel_group([[0]], 0, "gloo", "ep", use_all2all=True)
+    assert received["modern"] is True
+
+    parallel_state.init_vllm_model_parallel_group([[0]], 0, "gloo", "tp")
+    assert received["modern"] is False
+
+    def legacy(group_ranks, local_rank, backend, group_name=None, use_device_communicator=True):
+        received["legacy"] = True
+        return object()
+
+    monkeypatch.setattr(parallel_state.vllm_parallel_state, "init_model_parallel_group", legacy)
+    parallel_state.init_vllm_model_parallel_group([[0]], 0, "gloo", "ep", use_all2all=True)
+    assert received["legacy"] is True
@@ -110,7 +110,7 @@ class TestFusedMoEFactory:
        monkeypatch.setitem(
            sys.modules,
            "vllm.model_executor.layers.fused_moe",
-            SimpleNamespace(FusedMoE=mock_vllm_fused_moe),
+            SimpleNamespace(FusedMoEFactory=mock_vllm_fused_moe),
        )

        result = fused_moe.FusedMoE(prefix="test", a=1, reduce_results=True)
@@ -139,7 +139,7 @@ class TestFusedMoEFactory:
        monkeypatch.setitem(
            sys.modules,
            "vllm.model_executor.layers.fused_moe",
-            SimpleNamespace(FusedMoE=mocker.MagicMock(return_value=mock_runner)),
+            SimpleNamespace(FusedMoEFactory=mocker.MagicMock(return_value=mock_runner)),
        )
        mocker.patch.object(fused_moe, "current_omni_platform", mocker.MagicMock())
        mock_set_num_tokens = mocker.patch.object(fused_moe, "_set_forward_context_num_tokens")
@@ -169,7 +169,7 @@ class TestFusedMoEFactory:
        monkeypatch.setitem(
            sys.modules,
            "vllm.model_executor.layers.fused_moe",
-            SimpleNamespace(FusedMoE=mocker.MagicMock(return_value=mock_runner)),
+            SimpleNamespace(FusedMoEFactory=mocker.MagicMock(return_value=mock_runner)),
        )
        mocker.patch.object(fused_moe, "current_omni_platform", mocker.MagicMock())
        mock_set_num_tokens = mocker.patch.object(fused_moe, "_set_forward_context_num_tokens")
@@ -0,0 +1,438 @@
+from collections import defaultdict
+
+import pytest
+import torch
+import torch.nn as nn
+from diffusers.loaders.lora_conversion_utils import (
+    _convert_non_diffusers_qwen_lora_to_diffusers,
+)
+from pytest_mock import MockerFixture
+from torch.testing import assert_close
+
+from vllm_omni.diffusion.lora.loader import (
+    LoraLoaderMixin,
+    QwenImageLoraLoaderMixin,
+    WanLoraLoaderMixin,
+    _prepare_lora_delta,
+    _remap_state_dict_keys,
+)
+
+HEAD_DIM = 24
+RANK = 2
+NUM_LAYERS = 10
+
+pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
+
+
+class DummyTransformerBlock(nn.Module):
+    def __init__(
+        self,
+        d_model: int,
+        bias: bool = False,
+        stacked_params: bool = False,
+    ):
+        super().__init__()
+        self.attn = nn.Module()
+        if stacked_params:
+            self.attn.to_qkv = nn.Linear(d_model * 3, d_model, bias=bias)
+        else:
+            self.attn.to_q = nn.Linear(d_model, d_model, bias=bias)
+            self.attn.to_k = nn.Linear(d_model, d_model, bias=bias)
+            self.attn.to_v = nn.Linear(d_model, d_model, bias=bias)
+        self.attn.to_out = nn.Linear(d_model, d_model, bias=bias)
+
+
+class DummyTransformer(nn.Module):
+    def __init__(self, num_layers: int, d_model: int, bias: bool = False, stacked_params: bool = False):
+        super().__init__()
+
+        self.blocks = nn.ModuleList(
+            [DummyTransformerBlock(d_model, bias=bias, stacked_params=stacked_params) for _ in range(num_layers)]
+        )
+        if stacked_params:
+            self.stacked_params_mapping = [
+                (".attn.to_qkv", ".attn.to_q", "q"),
+                (".attn.to_qkv", ".attn.to_k", "k"),
+                (".attn.to_qkv", ".attn.to_v", "v"),
+            ]
+
+
+class DummyPipeline(nn.Module, LoraLoaderMixin):
+    def __init__(self, num_layers: int, d_model: int, bias: bool = False, stacked_params: bool = False):
+        super().__init__()
+        self.transformer = DummyTransformer(num_layers, d_model, bias=bias, stacked_params=stacked_params)
+
+
+# ======================================================
+# helper functions for test
+# ======================================================
+def _filter_parameters(named_params, mode: str = "all"):
+    if mode == "all":
+        return named_params
+
+    filtered = []
+    for name, param in named_params:
+        if name.endswith(mode):
+            filtered.append((name, param))
+
+    return filtered
+
+
+def assert_params_equal(module, original_params, mode: str = "all"):
+    for name, param in _filter_parameters(module.named_parameters(), mode):
+        assert_close(param, original_params[name])
+
+
+def assert_params_not_equal(module, original_params, mode: str = "all"):
+    for name, param in _filter_parameters(module.named_parameters(), mode):
+        assert not torch.allclose(param, original_params[name])
+
+
+def make_dummy_lora_state_dict(
+    base_key,
+    weight_a: torch.Tensor | None = None,
+    weight_b: torch.Tensor | None = None,
+    weight_bias: torch.Tensor | None = None,
+    bias: bool = False,
+    lora_a_suffix: str = "lora_A.weight",
+    lora_b_suffix: str = "lora_B.weight",
+    lora_bias_suffix: str = "bias",
+):
+    """
+    Create a dummy LoRA state dict.
+
+    Returns:
+        tuple: (state_dict, expected_delta)
+    """
+    key_a = f"{base_key}.{lora_a_suffix}"
+    key_b = f"{base_key}.{lora_b_suffix}"
+    key_bias = f"{base_key}.{lora_bias_suffix}" if bias else None
+
+    if weight_a is None:
+        weight_a = torch.randn(RANK, HEAD_DIM)
+    if weight_b is None:
+        weight_b = torch.randn(HEAD_DIM, RANK)
+
+    state_dict = {}
+    expected_delta = None
+
+    if bias:
+        state_dict[key_a] = weight_a
+        state_dict[key_b] = weight_b
+        weight_bias = torch.randn(HEAD_DIM)
+        state_dict[key_bias] = weight_bias
+        expected_delta = weight_bias
+    else:
+        state_dict[key_a] = weight_a
+        state_dict[key_b] = weight_b
+        expected_delta = weight_b @ weight_a
+
+    return state_dict, expected_delta
+
+
+def make_lora_state_dict_for_module(
+    module,
+    lora_rank=RANK,
+    bias=False,
+    lora_a_suffix: str = "lora_A.weight",
+    lora_b_suffix: str = "lora_B.weight",
+    lora_bias_suffix: str = "bias",
+    prefix: str = "transformer",
+    remap_proj_out: bool = True,
+):
+    param_to_weight_names = defaultdict(list)
+    if hasattr(module, "stacked_params_mapping"):
+        for param_name, weight_name, _ in module.stacked_params_mapping:
+            param_to_weight_names[param_name].append(weight_name)
+
+    state_dict = {}
+    for name, param in module.named_parameters(prefix=prefix):
+        if name.endswith(".weight"):
+            base_key = name[: -len(".weight")]
+            if remap_proj_out and base_key.endswith(".to_out"):
+                base_key = base_key.replace(".to_out", ".to_out.0")
+
+            d_in, d_out = param.shape
+            is_stacked_param = False
+            for param_name, weight_names in param_to_weight_names.items():
+                if param_name not in name:
+                    continue
+                is_stacked_param = True
+                target_in = d_in // len(weight_names)
+                for weight_name in weight_names:
+                    lora_key = base_key.replace(param_name, weight_name)
+                    weight_a = torch.randn(lora_rank, d_out)
+                    weight_b = torch.randn(target_in, lora_rank)
+                    state_dict[f"{lora_key}.{lora_a_suffix}"] = weight_a
+                    state_dict[f"{lora_key}.{lora_b_suffix}"] = weight_b
+
+            if is_stacked_param:
+                continue
+
+            weight_a = torch.randn(lora_rank, d_out)
+            weight_b = torch.randn(d_in, lora_rank)
+            key_a = f"{base_key}.{lora_a_suffix}"
+            key_b = f"{base_key}.{lora_b_suffix}"
+            state_dict[key_a] = weight_a
+            state_dict[key_b] = weight_b
+        elif name.endswith(".bias") and bias:
+            base_key = name[: -len(".bias")]
+            if remap_proj_out and base_key.endswith(".to_out"):
+                base_key = base_key.replace(".to_out", ".to_out.0")
+
+            d_bias = param.shape[0]
+            is_stacked_param = False
+            for param_name, weight_names in param_to_weight_names.items():
+                if param_name not in name:
+                    continue
+                is_stacked_param = True
+                target_bias = d_bias // len(weight_names)
+                for weight_name in weight_names:
+                    lora_key = base_key.replace(param_name, weight_name)
+                    weight_bias = torch.randn(target_bias)
+                    state_dict[f"{lora_key}.{lora_bias_suffix}"] = weight_bias
+
+            if is_stacked_param:
+                continue
+
+            key_bias = f"{base_key}.{lora_bias_suffix}"
+            state_dict[key_bias] = torch.randn_like(param)
+
+    return state_dict
+
+
+# ======================================================
+# Test loader helper functions
+# ======================================================
+def test_prepare_lora_delta_with_stacked_params():
+    base_key = "transformer.attn.to_qkv"
+
+    sd0, delta0 = make_dummy_lora_state_dict("transformer.attn.to_q")
+    sd1, delta1 = make_dummy_lora_state_dict("transformer.attn.to_k")
+    sd2, delta2 = make_dummy_lora_state_dict("transformer.attn.to_v")
+
+    lora_state_dict = {}
+    lora_state_dict.update(sd0)
+    lora_state_dict.update(sd1)
+    lora_state_dict.update(sd2)
+    expected_delta = torch.concat([delta0, delta1, delta2])
+
+    stacked_mapping = {".attn.to_qkv": [".attn.to_q", ".attn.to_k", ".attn.to_v"]}
+
+    actual_delta, used_keys = _prepare_lora_delta(
+        lora_state_dict,
+        base_key,
+        stacked_mapping,
+    )
+
+    # validate the delta
+    assert actual_delta is not None
+    assert_close(actual_delta, expected_delta)
+
+    # validate the used keys
+    key_q_a = "transformer.attn.to_q.lora_A.weight"
+    key_q_b = "transformer.attn.to_q.lora_B.weight"
+    key_k_a = "transformer.attn.to_k.lora_A.weight"
+    key_k_b = "transformer.attn.to_k.lora_B.weight"
+    key_v_a = "transformer.attn.to_v.lora_A.weight"
+    key_v_b = "transformer.attn.to_v.lora_B.weight"
+
+    for key in [key_q_a, key_q_b, key_k_a, key_k_b, key_v_a, key_v_b]:
+        assert key in used_keys
+
+
+def test_remap_state_dict_keys_matched_two():
+    lora_state_dict, _ = make_dummy_lora_state_dict(
+        "transformer.attn.to_out.0", bias=True, lora_bias_suffix="lora_B.bias"
+    )
+
+    lora_state_dict = _remap_state_dict_keys(
+        lora_state_dict,
+        [(".to_out.0.", ".to_out."), (".lora_B.bias", ".bias")],
+    )
+
+    assert len(lora_state_dict) == 3
+    assert "transformer.attn.to_out.bias" in lora_state_dict
+
+
+# ======================================================
+# Test LoraLoaderMixin
+# ======================================================
+class TestLoraLoaderMixin:
+    @pytest.mark.parametrize(
+        "bias, stacked_params",
+        [
+            (False, False),
+            (True, False),
+            (False, True),
+            (True, True),
+        ],
+        ids=[
+            "basic",
+            "bias",
+            "stacked_params",
+            "bias_stacked_params",
+        ],
+    )
+    def test_load_lora_into_module(self, bias: bool, stacked_params: bool):
+        pipeline = DummyPipeline(NUM_LAYERS, HEAD_DIM, bias=bias, stacked_params=stacked_params)
+        original_params = {name: param.clone() for name, param in pipeline.transformer.named_parameters()}
+
+        lora_state_dict = make_lora_state_dict_for_module(pipeline.transformer, bias=bias)
+        assert len(lora_state_dict) > 0
+
+        # remap the keys to vllm-omni format
+        lora_state_dict = _remap_state_dict_keys(lora_state_dict, [(".to_out.0.", ".to_out.")])
+
+        used_keys = pipeline.load_lora_into_module(lora_state_dict, pipeline.transformer)
+        # validate the parameters are updated
+        assert_params_not_equal(pipeline.transformer, original_params)
+
+        # validate all keys are used
+        missing_keys = set(lora_state_dict.keys()) - used_keys
+        assert len(missing_keys) == 0
+
+    def test_unload_module_lora(self):
+        pipeline = DummyPipeline(NUM_LAYERS, HEAD_DIM)
+        original_params = {name: param.clone() for name, param in pipeline.transformer.named_parameters()}
+
+        lora_state_dict = make_lora_state_dict_for_module(pipeline.transformer)
+        # validate the state dict is valid
+        assert len(lora_state_dict) > 0
+
+        # remap the keys to vllm-omni format
+        lora_state_dict = _remap_state_dict_keys(lora_state_dict, [(".to_out.0.", ".to_out.")])
+
+        used_keys = pipeline.load_lora_into_module(lora_state_dict, pipeline.transformer)
+
+        # validate the weights are updated
+        assert_params_not_equal(pipeline.transformer, original_params)
+
+        # validate all keys are used
+        missing_keys = set(lora_state_dict.keys()) - used_keys
+        assert len(missing_keys) == 0
+
+        unload_used_keys = pipeline.unload_module_lora(lora_state_dict, pipeline.transformer)
+        missing_keys = set(lora_state_dict.keys()) - unload_used_keys
+        assert len(missing_keys) == 0
+        assert len(unload_used_keys - used_keys) == 0
+
+        # validate the weights are restored
+        assert_params_equal(pipeline.transformer, original_params)
+
+
+# ======================================================
+# Test QwenImageLoraLoaderMixin
+# ======================================================
+class DummyQwenImagePipeline(nn.Module, QwenImageLoraLoaderMixin):
+    def __init__(self, num_layers: int, d_model: int, bias: bool = False):
+        super().__init__()
+        self.transformer = DummyTransformer(num_layers, d_model, bias=bias, stacked_params=True)
+
+
+class TestQwenImageLoraLoaderMixin:
+    def test_load_lora_weights(self, mocker: MockerFixture):
+        pipeline = DummyQwenImagePipeline(NUM_LAYERS, HEAD_DIM)
+        original_params = {name: param.clone() for name, param in pipeline.transformer.named_parameters()}
+
+        lora_state_dict = make_lora_state_dict_for_module(pipeline.transformer)
+        assert len(lora_state_dict) > 0
+
+        mocker.patch(
+            "vllm_omni.diffusion.lora.loader.get_converter_by_pipeline",
+            return_value=_convert_non_diffusers_qwen_lora_to_diffusers,
+        )
+        pipeline.load_lora_weights(lora_state_dict, "adapter0")
+
+        # validate the weights are updated after lora loaded
+        assert_params_not_equal(pipeline.transformer, original_params)
+
+        # validate lora_loaded map is updated after lora loaded
+        assert "adapter0" in pipeline.lora_loaded
+
+    def test_unload_lora_weights(self, mocker: MockerFixture):
+        pipeline = DummyQwenImagePipeline(NUM_LAYERS, HEAD_DIM)
+        original_params = {name: param.clone() for name, param in pipeline.transformer.named_parameters()}
+
+        lora_state_dict = make_lora_state_dict_for_module(pipeline.transformer)
+        assert len(lora_state_dict) > 0
+
+        mocker.patch(
+            "vllm_omni.diffusion.lora.loader.get_converter_by_pipeline",
+            return_value=_convert_non_diffusers_qwen_lora_to_diffusers,
+        )
+        pipeline.load_lora_weights(lora_state_dict, "adapter0")
+        # validate the weights are updated after lora loaded
+        assert_params_not_equal(pipeline.transformer, original_params)
+
+        # validate lora_loaded map is updated after lora loaded
+        assert "adapter0" in pipeline.lora_loaded
+
+        pipeline.unload_lora_weights("adapter0")
+        # validate the weights are restored after lora unloaded
+        assert_params_equal(pipeline.transformer, original_params)
+
+        # validate lora_loaded map is updated
+        assert "adapter0" not in pipeline.lora_loaded
+
+    def test_load_diffusers_lora_applies_non_unit_alpha(self, mocker: MockerFixture):
+        pipeline = DummyQwenImagePipeline(1, HEAD_DIM)
+        to_out = pipeline.transformer.blocks[0].attn.to_out
+        original_weight = to_out.weight.detach().clone()
+        base_key = "blocks.0.attn.to_out.0"
+        lora_a = torch.randn(RANK, HEAD_DIM)
+        lora_b = torch.randn(HEAD_DIM, RANK)
+        alpha = torch.tensor(6.0)
+        state_dict = {
+            f"{base_key}.lora_A.weight": lora_a,
+            f"{base_key}.lora_B.weight": lora_b,
+            f"{base_key}.alpha": alpha,
+        }
+        mocker.patch(
+            "vllm_omni.diffusion.lora.loader.get_converter_by_pipeline",
+            return_value=_convert_non_diffusers_qwen_lora_to_diffusers,
+        )
+
+        pipeline.load_lora_weights(state_dict, "adapter0")
+
+        expected_weight = original_weight + (lora_b @ lora_a) * (alpha.item() / RANK)
+        assert_close(to_out.weight, expected_weight)
+
+        pipeline.unload_lora_weights("adapter0")
+        assert_close(to_out.weight, original_weight)
+
+
+# ======================================================
+# Test WanLoraLoaderMixin
+# ======================================================
+class DummyWanPipeline(nn.Module, WanLoraLoaderMixin):
+    def __init__(self, num_layers: int, d_model: int):
+        super().__init__()
+        self.has_transformer_2 = True
+        self.transformer = DummyTransformer(num_layers, d_model, stacked_params=True)
+        self.transformer_2 = DummyTransformer(num_layers, d_model, stacked_params=True)
+
+
+def test_wan_load_lora_weights_targets_both_transformers(mocker: MockerFixture):
+    pipeline = DummyWanPipeline(NUM_LAYERS, HEAD_DIM)
+    original_high = {name: param.clone() for name, param in pipeline.transformer.named_parameters()}
+    original_low = {name: param.clone() for name, param in pipeline.transformer_2.named_parameters()}
+    high_state_dict = make_lora_state_dict_for_module(pipeline.transformer)
+    low_state_dict = make_lora_state_dict_for_module(pipeline.transformer_2)
+
+    mocker.patch(
+        "vllm_omni.diffusion.lora.loader._load_lora_state_dict",
+        side_effect=[high_state_dict, low_state_dict],
+    )
+
+    pipeline.load_lora_weights(["high.safetensors", "low.safetensors"], "adapter0")
+
+    assert_params_not_equal(pipeline.transformer, original_high)
+    assert_params_not_equal(pipeline.transformer_2, original_low)
+    assert set(pipeline.lora_loaded["adapter0"]) == {"transformer", "transformer_2"}
+
+    pipeline.unload_lora_weights("adapter0")
+
+    assert_params_equal(pipeline.transformer, original_high)
+    assert_params_equal(pipeline.transformer_2, original_low)
+    assert "adapter0" not in pipeline.lora_loaded
@@ -0,0 +1,447 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+from types import SimpleNamespace
+
+import pytest
+import torch
+from transformers.utils.generic import ModelOutput
+
+import vllm_omni.diffusion.models.hunyuan_image3.pipeline_hunyuan_image3 as pipeline_module
+from vllm_omni.diffusion.diffusion_kv.config import DiffusionKVCacheMode
+from vllm_omni.diffusion.diffusion_kv.request import DiffusionKVRequest
+from vllm_omni.diffusion.models.hunyuan_image3.hunyuan_image3_tokenizer import TokenizerEncodeOutput
+from vllm_omni.diffusion.models.hunyuan_image3.hunyuan_image3_transformer import (
+    ImageInfo,
+    JointImageInfo,
+)
+from vllm_omni.diffusion.models.hunyuan_image3.pipeline_hunyuan_image3 import (
+    HunyuanImage3Pipeline,
+    get_hunyuan_image_3_pre_process_func,
+)
+from vllm_omni.diffusion.models.hunyuan_image3.request_layout import (
+    HunyuanPreparedLayout,
+    build_hunyuan_diffusion_kv_requests,
+    extract_hunyuan_prompt_inputs,
+    hunyuan_num_image_tokens,
+    normalize_hunyuan_cot_text,
+    prepare_hunyuan_layout,
+)
+from vllm_omni.diffusion.request import OmniDiffusionRequest
+from vllm_omni.inputs.data import OmniDiffusionSamplingParams
+
+pytestmark = [pytest.mark.core_model, pytest.mark.cpu, pytest.mark.diffusion]
+
+
+def test_dense_legacy_preprocess_does_not_initialize_layout_tokenizer(monkeypatch) -> None:
+    hf_config = SimpleNamespace(vae_downsample_factor=(8, 8), patch_size=2)
+    image_processor = SimpleNamespace(vision_encoder_processor=SimpleNamespace(patch_size=1))
+    monkeypatch.setattr(pipeline_module, "get_config", lambda *_args, **_kwargs: hf_config)
+    monkeypatch.setattr(pipeline_module, "HunyuanImage3ImageProcessor", lambda _config: image_processor)
+    monkeypatch.setattr(
+        pipeline_module,
+        "TokenizerWrapper",
+        lambda _model: pytest.fail("dense_legacy must not initialize the Diffusion KV tokenizer"),
+    )
+
+    get_hunyuan_image_3_pre_process_func(
+        SimpleNamespace(model="model", diffusion_kv_mode=DiffusionKVCacheMode.DENSE_LEGACY)
+    )
+
+
+class _FakeTokenizerWrapper:
+    def __init__(self, *, prefix_lens: list[int]) -> None:
+        self.prefix_lens = prefix_lens
+        self.calls: list[dict] = []
+        self.pad_token_id = 0
+        self.eos_token_id = 2
+        self.boi_token_id = 3
+        self.end_recaption_token_id = 4
+        self.end_answer_token_id = 5
+        self.special_token_map = {f"<img_ratio_{index}>": 100 + index for index in range(33)}
+
+    def apply_chat_template(self, **kwargs):
+        self.calls.append(kwargs)
+        rows = len(self.prefix_lens)
+        has_cond_image = kwargs.get("batch_cond_image_info") is not None
+        seq_lens = [prefix_len + 20 for prefix_len in self.prefix_lens]
+        padded_seq_len = max(seq_lens)
+        positions = torch.arange(padded_seq_len)
+        joint_image_slices = [[slice(1, 3), slice(3, 5)] if has_cond_image else [] for _ in range(rows)]
+        gen_image_slices = [[slice(length + 1, length + 17)] for length in self.prefix_lens]
+        all_image_slices = [joint + generated for joint, generated in zip(joint_image_slices, gen_image_slices)]
+        cond_vae_image_mask = (
+            torch.stack([(positions >= 1) & (positions < 3) for _ in range(rows)]) if has_cond_image else None
+        )
+        cond_vit_image_mask = (
+            torch.stack([(positions >= 3) & (positions < 5) for _ in range(rows)]) if has_cond_image else None
+        )
+        sections = []
+        for _ in range(rows):
+            row_sections = []
+            if has_cond_image:
+                row_sections.append(
+                    {
+                        "type": "joint_image",
+                        "token_height": [1, 1],
+                        "token_width": [2, 2],
+                    }
+                )
+            row_sections.append({"type": "gen_image", "token_height": 4, "token_width": 4})
+            sections.append(row_sections)
+        return {
+            "output": TokenizerEncodeOutput(
+                tokens=torch.arange(rows * padded_seq_len, dtype=torch.long).reshape(rows, padded_seq_len),
+                gen_timestep_scatter_index=torch.tensor(self.prefix_lens, dtype=torch.long).reshape(rows, 1),
+                cond_timestep_scatter_index=(torch.zeros(rows, 1, dtype=torch.long) if has_cond_image else None),
+                all_image_slices=all_image_slices,
+                gen_image_mask=torch.stack(
+                    [(positions >= length + 1) & (positions < length + 17) for length in self.prefix_lens]
+                ),
+                cond_vae_image_mask=cond_vae_image_mask,
+                cond_vit_image_mask=cond_vit_image_mask,
+                joint_image_slices=joint_image_slices,
+                gen_image_slices=gen_image_slices,
+                real_pos=torch.tensor(seq_lens, dtype=torch.long).reshape(rows, 1),
+            ),
+            "sections": sections,
+        }
+
+
+class _FakeImageProcessor:
+    def __init__(self, image_token_length: int = 16) -> None:
+        self.image_token_length = image_token_length
+        self.image_sizes: list[tuple[int, int]] = []
+        self.vision_encoder_processor = SimpleNamespace(patch_size=1)
+
+    def build_image_info(self, image_size):
+        self.image_sizes.append(image_size)
+        return ImageInfo(
+            image_type="gen_image",
+            image_width=image_size[1],
+            image_height=image_size[0],
+            token_width=4,
+            token_height=4,
+            image_token_length=self.image_token_length,
+            base_size=1024,
+            ratio_index=0,
+        )
+
+
+def _components(prefix_lens: list[int]):
+    return _FakeTokenizerWrapper(prefix_lens=prefix_lens), _FakeImageProcessor()
+
+
+def _prepare(
+    request: OmniDiffusionRequest,
+    tokenizer: _FakeTokenizerWrapper,
+    image_processor: _FakeImageProcessor,
+) -> HunyuanPreparedLayout:
+    prepared_layout = prepare_hunyuan_layout(
+        request,
+        tokenizer_wrapper=tokenizer,
+        image_processor=image_processor,
+        generation_config=SimpleNamespace(sequence_template="instruct", drop_think=False),
+        image_base_size=1024,
+    )
+    request.prepared_layout = prepared_layout
+    return prepared_layout
+
+
+def _request(
+    *,
+    guidance_scale: float,
+    prompt="draw a cat",
+    request_id: str = "req",
+) -> OmniDiffusionRequest:
+    return OmniDiffusionRequest(
+        prompt=prompt,
+        sampling_params=OmniDiffusionSamplingParams(
+            height=768,
+            width=1024,
+            guidance_scale=guidance_scale,
+            num_inference_steps=4,
+        ),
+        request_id=request_id,
+    )
+
+
+def test_builds_kv_request_lengths_without_model_execution() -> None:
+    tokenizer, image_processor = _components([12])
+    request = _request(guidance_scale=1.0)
+    prepared_layout = _prepare(request, tokenizer, image_processor)
+
+    kv_requests = build_hunyuan_diffusion_kv_requests(request, prepared_layout)
+
+    assert isinstance(request.prepared_layout, HunyuanPreparedLayout)
+    assert all(isinstance(item, DiffusionKVRequest) for item in kv_requests)
+    assert [(item.prefix_len, item.target_len, item.seq_len) for item in kv_requests] == [(12, 17, 32)]
+    assert image_processor.image_sizes == [(768, 1024)]
+    assert tokenizer.calls[0]["sequence_template"] == "instruct"
+    assert tokenizer.calls[0]["cfg_factor"] == 1
+    assert kv_requests[0].block_hashes == []
+    assert kv_requests[0].kv_contexts == ()
+    assert kv_requests[0].skip_reading_prefix_cache is True
+
+
+def test_builds_one_kv_request_per_cfg_row() -> None:
+    tokenizer, image_processor = _components([12, 14])
+    request = _request(guidance_scale=5.0)
+    prepared_layout = _prepare(request, tokenizer, image_processor)
+
+    kv_requests = build_hunyuan_diffusion_kv_requests(request, prepared_layout)
+
+    assert [item.sequence_id for item in kv_requests] == [0, 1]
+    assert [item.prefix_len for item in kv_requests] == [12, 14]
+    assert [item.seq_len for item in kv_requests] == [32, 34]
+    assert tokenizer.calls[0]["cfg_factor"] == 2
+
+
+def _reference_image() -> JointImageInfo:
+    return JointImageInfo(
+        vae_image_info=ImageInfo(
+            image_type="vae",
+            token_width=8,
+            token_height=8,
+            image_token_length=64,
+        ),
+        vision_image_info=ImageInfo(
+            image_type="siglip2",
+            token_width=4,
+            token_height=4,
+            image_token_length=16,
+        ),
+        vision_encoder_kwargs={
+            "spatial_shapes": torch.tensor([4, 4]),
+            "pixel_attention_mask": torch.ones(4, dtype=torch.bool),
+        },
+    )
+
+
+def test_passes_preprocessed_reference_image_geometry_to_tokenizer() -> None:
+    tokenizer, image_processor = _components([20])
+    joint_image = _reference_image()
+    prompt = {
+        "prompt": "edit this image",
+        "additional_information": {"batch_cond_image_info": [joint_image]},
+    }
+    request = _request(guidance_scale=1.0, prompt=prompt)
+    prepared_layout = _prepare(request, tokenizer, image_processor)
+
+    kv_requests = build_hunyuan_diffusion_kv_requests(request, prepared_layout)
+
+    assert tokenizer.calls[0]["batch_cond_image_info"] == [[joint_image]]
+    assert kv_requests[0].prefix_len == 20
+    assert kv_requests[0].target_len == 17
+    assert kv_requests[0].seq_len == 40
+    assert kv_requests[0].kv_contexts == ()
+
+
+def _assert_nested_equal(actual, expected) -> None:
+    if isinstance(expected, torch.Tensor):
+        torch.testing.assert_close(actual, expected)
+    elif isinstance(expected, dict):
+        assert actual.keys() == expected.keys()
+        for key in expected:
+            _assert_nested_equal(actual[key], expected[key])
+    elif isinstance(expected, list | tuple):
+        assert isinstance(actual, type(expected))
+        assert len(actual) == len(expected)
+        for actual_item, expected_item in zip(actual, expected):
+            _assert_nested_equal(actual_item, expected_item)
+    else:
+        assert actual == expected
+
+
+@pytest.mark.parametrize(
+    ("prefix_lens", "guidance_scale", "with_reference_image"),
+    [
+        pytest.param([12], 1.0, False, id="cfg-1"),
+        pytest.param([12, 14], 5.0, False, id="cfg-2"),
+        pytest.param([20], 1.0, True, id="cfg-1-reference-image"),
+        pytest.param([20, 22], 5.0, True, id="cfg-2-reference-image"),
+    ],
+)
+def test_prepared_model_inputs_match_local_tokenization(
+    monkeypatch,
+    prefix_lens: list[int],
+    guidance_scale: float,
+    with_reference_image: bool,
+) -> None:
+    tokenizer, image_processor = _components(prefix_lens)
+    prompt: str | dict = "draw a cat"
+    if with_reference_image:
+        prompt = {
+            "prompt": "edit this image",
+            "additional_information": {"batch_cond_image_info": [_reference_image()]},
+        }
+    request = _request(guidance_scale=guidance_scale, prompt=prompt)
+    prepared_layout = _prepare(request, tokenizer, image_processor)
+
+    prompts, cot_texts, system_prompt, batch_cond_image_info, bot_task = extract_hunyuan_prompt_inputs(
+        [request.prompt],
+        request.sampling_params.extra_args or {},
+        request_id=request.request_id,
+        allow_cond_image=True,
+    )
+    cot_text = (
+        [normalize_hunyuan_cot_text(text) for text in cot_texts]
+        if any(text is not None for text in cot_texts)
+        else None
+    )
+
+    pipeline = object.__new__(HunyuanImage3Pipeline)
+    pipeline._tkwrapper = tokenizer
+    pipeline.image_processor = image_processor
+    pipeline.generation_config = SimpleNamespace(sequence_template="instruct", drop_think=False)
+    pipeline.config = SimpleNamespace(
+        image_base_size=1024,
+        attention_head_dim=2,
+        rope_theta=10000.0,
+    )
+    monkeypatch.setattr(HunyuanImage3Pipeline, "device", property(lambda _self: torch.device("cpu")))
+
+    def fake_encode_cond_image(batch_cond_image_info_list, cfg_factor=1, generator=None):
+        del generator
+        rows = len(batch_cond_image_info_list) * cfg_factor
+        return (
+            torch.arange(rows * 4, dtype=torch.float32).reshape(rows, 1, 2, 2),
+            torch.arange(rows, dtype=torch.float32),
+            torch.arange(rows * 3, dtype=torch.float32).reshape(rows, 3),
+        )
+
+    pipeline._encode_cond_image = fake_encode_cond_image
+
+    def fake_build_batch_2d_rope(*, image_infos, seq_len, **_kwargs):
+        geometry_sum = sum(
+            int(image_slice.start or 0) + int(image_slice.stop or 0) + height + width
+            for row in image_infos
+            for image_slice, (height, width) in row
+        )
+        values = torch.tensor([len(image_infos), seq_len, geometry_sum], dtype=torch.float32)
+        return values, values + 1
+
+    monkeypatch.setattr(pipeline_module, "build_batch_2d_rope", fake_build_batch_2d_rope)
+    common_kwargs = dict(
+        prompt=prompts,
+        cot_text=cot_text,
+        system_prompt=system_prompt,
+        mode="gen_image",
+        guidance_scale=guidance_scale,
+        image_size=(request.sampling_params.height, request.sampling_params.width),
+        generator=[torch.Generator().manual_seed(0)],
+        batch_cond_image_info=batch_cond_image_info,
+        bot_task=bot_task,
+    )
+
+    local_inputs = pipeline.prepare_model_inputs(**common_kwargs)
+    prepared_inputs = pipeline.prepare_model_inputs(**common_kwargs, prepared_layout=prepared_layout)
+
+    comparable_fields = (
+        "input_ids",
+        "position_ids",
+        "custom_pos_emb",
+        "image_mask",
+        "gen_timestep_scatter_index",
+        "cond_vae_images",
+        "cond_timestep",
+        "cond_vae_image_mask",
+        "cond_vit_images",
+        "cond_vit_image_mask",
+        "vit_kwargs",
+        "cond_timestep_scatter_index",
+    )
+    for field in comparable_fields:
+        _assert_nested_equal(prepared_inputs[field], local_inputs[field])
+
+    tokenizer_fields = (
+        "tokens",
+        "gen_image_mask",
+        "cond_vae_image_mask",
+        "cond_vit_image_mask",
+        "gen_timestep_scatter_index",
+        "cond_timestep_scatter_index",
+        "real_pos",
+        "all_image_slices",
+        "joint_image_slices",
+        "gen_image_slices",
+    )
+    for field in tokenizer_fields:
+        _assert_nested_equal(
+            getattr(prepared_inputs["tokenizer_output"], field),
+            getattr(local_inputs["tokenizer_output"], field),
+        )
+
+    local_mask = pipeline._prepare_attention_mask_for_generation(
+        local_inputs["input_ids"],
+        pipeline.generation_config,
+        local_inputs,
+    )
+    prepared_mask = pipeline._prepare_attention_mask_for_generation(
+        prepared_inputs["input_ids"],
+        pipeline.generation_config,
+        prepared_inputs,
+    )
+    torch.testing.assert_close(prepared_mask, local_mask)
+    assert prepared_inputs["full_attn_spans"] == local_inputs["full_attn_spans"]
+
+    num_image_tokens = hunyuan_num_image_tokens(prepared_layout.generated_image_info)
+    local_inputs.update(attention_mask=local_mask, num_image_tokens=num_image_tokens)
+    prepared_inputs.update(attention_mask=prepared_mask, num_image_tokens=num_image_tokens)
+    local_step_inputs = pipeline._update_model_kwargs_for_generation(ModelOutput(), local_inputs)
+    prepared_step_inputs = pipeline._update_model_kwargs_for_generation(ModelOutput(), prepared_inputs)
+
+    for field in ("position_ids", "attention_mask", "gen_timestep_scatter_index", "full_attn_spans"):
+        _assert_nested_equal(prepared_step_inputs[field], local_step_inputs[field])
+
+
+def test_rejects_tokenizer_cfg_row_mismatch() -> None:
+    tokenizer, image_processor = _components([12])
+    request = _request(guidance_scale=5.0)
+    prepared_layout = _prepare(request, tokenizer, image_processor)
+
+    with pytest.raises(ValueError, match="sequence count does not match"):
+        build_hunyuan_diffusion_kv_requests(request, prepared_layout)
+
+
+def test_sender_endpoint_does_not_change_local_kv_request() -> None:
+    tokenizer, image_processor = _components([12])
+    request = _request(guidance_scale=1.0)
+    request.kv_sender_info = {"host": "127.0.0.1", "zmq_port": 5000}
+    prepared_layout = _prepare(request, tokenizer, image_processor)
+
+    kv_requests = build_hunyuan_diffusion_kv_requests(request, prepared_layout)
+
+    assert kv_requests[0].seq_len == 32
+    assert kv_requests[0].num_computed_tokens == 0
+
+
+def test_paged_preprocess_attaches_layout_and_scheduler_kv_requests(monkeypatch) -> None:
+    tokenizer, image_processor = _components([12])
+    hf_config = SimpleNamespace(
+        vae_downsample_factor=(8, 8),
+        patch_size=2,
+        image_base_size=1024,
+    )
+    monkeypatch.setattr(pipeline_module, "get_config", lambda *_args, **_kwargs: hf_config)
+    monkeypatch.setattr(pipeline_module, "HunyuanImage3ImageProcessor", lambda _config: image_processor)
+    monkeypatch.setattr(pipeline_module, "TokenizerWrapper", lambda _model: tokenizer)
+    monkeypatch.setattr(
+        pipeline_module.GenerationConfig,
+        "from_pretrained",
+        lambda _model: SimpleNamespace(sequence_template="instruct", drop_think=False),
+    )
+    preprocess = get_hunyuan_image_3_pre_process_func(
+        SimpleNamespace(model="model", diffusion_kv_mode=DiffusionKVCacheMode.PAGED_SCHEDULER)
+    )
+    request = _request(guidance_scale=1.0)
+
+    prepared_request = preprocess(request)
+
+    assert prepared_request is request
+    assert isinstance(request.prepared_layout, HunyuanPreparedLayout)
+    assert request.diffusion_kv_requests is not None
+    assert [(item.prefix_len, item.target_len, item.seq_len) for item in request.diffusion_kv_requests] == [
+        (12, 17, 32)
+    ]
+    assert len(tokenizer.calls) == 1
@@ -7,7 +7,10 @@ import pytest
 import torch

 import vllm_omni.diffusion.models.hunyuan_image3.pipeline_hunyuan_image3 as hy3_module
+import vllm_omni.diffusion.models.hunyuan_image3.request_layout as hy3_layout_module
 from vllm_omni.diffusion.data import AttentionConfig, AttentionSpec
+from vllm_omni.diffusion.models.hunyuan_image3.hunyuan_image3_tokenizer import TokenizerEncodeOutput
+from vllm_omni.diffusion.models.hunyuan_image3.hunyuan_image3_transformer import ImageInfo
 from vllm_omni.diffusion.models.hunyuan_image3.pipeline_hunyuan_image3 import (
    _STEP_AR_KV,
    _STEP_CFG_FACTOR,
@@ -18,6 +21,7 @@ from vllm_omni.diffusion.models.hunyuan_image3.pipeline_hunyuan_image3 import (
    _STEP_PROMPT_KV,
    HunyuanImage3Pipeline,
 )
+from vllm_omni.diffusion.models.hunyuan_image3.request_layout import HunyuanPreparedLayout
 from vllm_omni.diffusion.worker.input_batch import InputBatch
 from vllm_omni.diffusion.worker.request_batch import DiffusionRequestBatch
 from vllm_omni.diffusion.worker.utils import StepRequestState
@@ -76,6 +80,72 @@ def _sampling_params(**extra_args):
    )


+def _prepared_layout() -> HunyuanPreparedLayout:
+    return HunyuanPreparedLayout(
+        tokenizer_output=TokenizerEncodeOutput(
+            tokens=torch.arange(21).reshape(1, 21),
+            gen_image_mask=torch.ones(1, 21, dtype=torch.bool),
+            gen_timestep_scatter_index=torch.tensor([[4]]),
+            all_image_slices=[[slice(5, 21)]],
+            joint_image_slices=[[]],
+            gen_image_slices=[[slice(5, 21)]],
+            real_pos=torch.tensor([[21]]),
+        ),
+        rope_image_info=[[(slice(5, 21), (4, 4))]],
+        generated_image_info=ImageInfo(
+            image_type="gen_image",
+            image_width=512,
+            image_height=512,
+            token_width=4,
+            token_height=4,
+            image_token_length=16,
+        ),
+    )
+
+
+def test_prepare_model_inputs_reuses_prepared_layout(monkeypatch):
+    pipeline = _pipeline()
+    monkeypatch.setattr(HunyuanImage3Pipeline, "device", property(lambda self: torch.device("cpu")))
+    rope_kwargs = {}
+
+    def fake_build_batch_2d_rope(**kwargs):
+        rope_kwargs.update(kwargs)
+        return torch.zeros(1), torch.zeros(1)
+
+    monkeypatch.setattr(hy3_module, "build_batch_2d_rope", fake_build_batch_2d_rope)
+
+    def fail_apply_chat_template(**_kwargs):
+        pytest.fail("prepared Hunyuan execution must not apply the chat template again")
+
+    pipeline._tkwrapper = SimpleNamespace(
+        apply_chat_template=fail_apply_chat_template,
+        eos_token_id=2,
+        boi_token_id=3,
+        end_recaption_token_id=4,
+        end_answer_token_id=5,
+    )
+    pipeline.config = SimpleNamespace(
+        attention_head_dim=2,
+        rope_theta=10000.0,
+    )
+    prepared = _prepared_layout()
+
+    model_inputs = pipeline.prepare_model_inputs(
+        prompt="prompt",
+        mode="gen_image",
+        guidance_scale=1.0,
+        image_size=(512, 512),
+        device=torch.device("cpu"),
+        generator=[torch.Generator().manual_seed(0)],
+        prepared_layout=prepared,
+    )
+
+    assert model_inputs["tokenizer_output"] is prepared.tokenizer_output
+    assert model_inputs["batch_gen_image_info"] == [prepared.generated_image_info]
+    assert rope_kwargs["image_infos"] is prepared.rope_image_info
+    torch.testing.assert_close(model_inputs["input_ids"], prepared.tokenizer_output.tokens)
+
+
 def test_hunyuan_step_group_key_ignores_step_index_for_later_steps():
    pipeline = _pipeline()
    states = [_state("req-0", 1), _state("req-1", 3)]
@@ -131,12 +201,13 @@ def test_prepare_encode_preserves_normal_hunyuan_bot_task_semantics(
        captured.update(kwargs)
        raise RuntimeError("stop after prepare_model_inputs")

-    monkeypatch.setattr(hy3_module, "get_system_prompt", fake_get_system_prompt)
+    monkeypatch.setattr(hy3_layout_module, "get_system_prompt", fake_get_system_prompt)
    pipeline.prepare_model_inputs = fake_prepare_model_inputs
    state = StepRequestState(
        request_id="req-bot-task",
        sampling=sampling,
        prompt=prompt_item,
+        prepared_layout=_prepared_layout(),
    )

    with pytest.raises(RuntimeError, match="stop after prepare_model_inputs"):
@@ -144,6 +215,7 @@ def test_prepare_encode_preserves_normal_hunyuan_bot_task_semantics(

    assert captured["bot_task"] == expected_model_bot_task
    assert captured["system_prompt_bot_task"] == expected_system_bot_task
+    assert captured["prepared_layout"] is state.prepared_layout


 def test_forward_uses_same_hunyuan_bot_task_semantics(monkeypatch):
@@ -159,23 +231,22 @@ def test_forward_uses_same_hunyuan_bot_task_semantics(monkeypatch):
        captured.update(kwargs)
        raise RuntimeError("stop after prepare_model_inputs")

-    monkeypatch.setattr(hy3_module, "get_system_prompt", fake_get_system_prompt)
+    monkeypatch.setattr(hy3_layout_module, "get_system_prompt", fake_get_system_prompt)
    pipeline.prepare_model_inputs = fake_prepare_model_inputs
-    req = DiffusionRequestBatch(
-        requests=[
-            SimpleNamespace(
-                request_id="req-forward-bot-task",
-                sampling_params=_sampling_params(bot_task="think_recaption", use_system_prompt="dynamic"),
-                prompt={"prompt": "prompt", "bot_task": "vanilla"},
-            )
-        ]
+    request = SimpleNamespace(
+        request_id="req-forward-bot-task",
+        sampling_params=_sampling_params(bot_task="think_recaption", use_system_prompt="dynamic"),
+        prompt={"prompt": "prompt", "bot_task": "vanilla"},
+        prepared_layout=_prepared_layout(),
    )
+    req = DiffusionRequestBatch(requests=[request])

    with pytest.raises(RuntimeError, match="stop after prepare_model_inputs"):
        pipeline.forward(req)

    assert captured["bot_task"] == "think"
    assert captured["system_prompt_bot_task"] == "think"
+    assert captured["prepared_layout"] is request.prepared_layout


 def test_grouped_denoise_rejects_non_sdpa_attention_backend():
@@ -0,0 +1,133 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+from types import SimpleNamespace
+
+import numpy as np
+import pytest
+import torch
+from PIL import Image
+from torch import nn
+
+from vllm_omni.diffusion.models.lingbot_video.image_condition import (
+    apply_clean_prefix,
+    geometry_align_image,
+    prepare_ti2v_image_condition,
+)
+
+pytestmark = [pytest.mark.core_model, pytest.mark.cpu, pytest.mark.diffusion]
+
+
+def _gradient_image(width: int, height: int) -> Image.Image:
+    x = np.linspace(0, 255, width, dtype=np.uint8)[None, :]
+    y = np.linspace(0, 255, height, dtype=np.uint8)[:, None]
+    rgb = np.stack(
+        [
+            np.broadcast_to(x, (height, width)),
+            np.broadcast_to(y, (height, width)),
+            np.full((height, width), 127, dtype=np.uint8),
+        ],
+        axis=-1,
+    )
+    return Image.fromarray(rgb, mode="RGB")
+
+
+@pytest.mark.parametrize("source_size", [(160, 64), (64, 160), (96, 96)])
+def test_geometry_align_image_supports_source_aspect_ratios(source_size):
+    pixel = geometry_align_image(_gradient_image(*source_size), 64, 96)
+
+    assert pixel.shape == (1, 3, 1, 64, 96)
+    assert pixel.dtype == torch.float32
+    assert 0 <= pixel.min() <= pixel.max() <= 1
+
+
+@pytest.mark.parametrize(
+    ("source_size", "target_size"),
+    [((160, 64), (64, 96)), ((64, 160), (96, 64))],
+)
+def test_geometry_align_image_center_crops_without_stretching(source_size, target_size):
+    width, height = source_size
+    target_height, target_width = target_size
+    rgb = np.full((height, width, 3), (255, 0, 0), dtype=np.uint8)
+    if width > target_width:
+        left = (width - target_width) // 2
+        rgb[:, left : left + target_width] = (0, 255, 0)
+    else:
+        top = (height - target_height) // 2
+        rgb[top : top + target_height] = (0, 255, 0)
+
+    pixel = geometry_align_image(
+        Image.fromarray(rgb, mode="RGB"),
+        target_height,
+        target_width,
+    )
+
+    assert pixel.shape == (1, 3, 1, target_height, target_width)
+    expected_green = torch.tensor((0.0, 1.0, 0.0)).view(1, 3, 1, 1, 1)
+    assert torch.equal(pixel, expected_green.expand_as(pixel))
+
+
+class _LatentDistribution:
+    def __init__(self, latent: torch.Tensor):
+        self.latent = latent
+
+    def sample(self, generator=None):
+        del generator
+        return self.latent
+
+
+class _CapturingVAE(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.anchor = nn.Parameter(torch.zeros(()))
+        self.config = SimpleNamespace(
+            latents_mean=[0.0, 0.0, 0.0],
+            latents_std=[1.0, 1.0, 1.0],
+        )
+        self.encoded_pixel = None
+
+    def encode(self, pixel):
+        self.encoded_pixel = pixel.detach().clone()
+        return SimpleNamespace(latent_dist=_LatentDistribution(pixel))
+
+
+def test_ti2v_vlm_and_vae_paths_share_geometry_aligned_image():
+    vae = _CapturingVAE()
+    condition = prepare_ti2v_image_condition(
+        _gradient_image(160, 64),
+        height=64,
+        width=96,
+        vae=vae,
+        vision_patch_size=16,
+        device=torch.device("cpu"),
+    )
+
+    vae_pixel = (vae.encoded_pixel + 1.0) / 2.0
+    vae_image = vae_pixel[0, :, 0].permute(1, 2, 0).mul(255).byte().numpy()
+    assert condition.vlm_image.size == (96, 64)
+    np.testing.assert_allclose(
+        np.asarray(condition.vlm_image),
+        vae_image,
+        rtol=0,
+        atol=1,
+    )
+
+
+def test_apply_clean_prefix_only_overwrites_condition_frames():
+    latents = torch.arange(20, dtype=torch.float32).reshape(1, 1, 5, 2, 2)
+    original_tail = latents[:, :, 2:].clone()
+    clean = torch.full((1, 1, 2, 2, 2), -3.0)
+
+    returned = apply_clean_prefix(latents, clean)
+
+    assert returned is latents
+    assert torch.equal(latents[:, :, :2], clean)
+    assert torch.equal(latents[:, :, 2:], original_tail)
+
+
+def test_apply_clean_prefix_rejects_incompatible_shapes():
+    with pytest.raises(ValueError, match="shape must match"):
+        apply_clean_prefix(
+            torch.zeros(1, 2, 3, 4, 4),
+            torch.zeros(1, 3, 1, 4, 4),
+        )
@@ -6,9 +6,10 @@ from types import SimpleNamespace
 import numpy as np
 import pytest
 import torch
+from PIL import Image
 from torch import nn

-from vllm_omni.diffusion.request import OmniDiffusionRequest
+from vllm_omni.diffusion.request import DUMMY_DIFFUSION_REQUEST_ID, OmniDiffusionRequest
 from vllm_omni.diffusion.worker.request_batch import DiffusionRequestBatch
 from vllm_omni.inputs.data import OmniDiffusionSamplingParams

@@ -22,6 +23,8 @@ def _make_pipeline():
    nn.Module.__init__(pipeline)
    pipeline.device = torch.device("cpu")
    pipeline.default_negative_prompt = "default negative"
+    pipeline.default_image_negative_prompt = "default image negative"
+    pipeline.img_prompt_template = "<image>"
    pipeline.od_config = SimpleNamespace(flow_shift=None)
    return pipeline

@@ -37,17 +40,24 @@ def test_lingbot_video_pipeline_import_and_registry():
        LingBotVideoPipeline,
        LingBotVideoTransformer3DModel,
        get_lingbot_video_post_process_func,
+        get_lingbot_video_pre_process_func,
+    )
+    from vllm_omni.diffusion.registry import (
+        _DIFFUSION_MODELS,
+        _DIFFUSION_POST_PROCESS_FUNCS,
+        _DIFFUSION_PRE_PROCESS_FUNCS,
    )
-    from vllm_omni.diffusion.registry import _DIFFUSION_MODELS, _DIFFUSION_POST_PROCESS_FUNCS

    assert LingBotVideoPipeline is not None
    assert LingBotVideoTransformer3DModel is not None
+    assert get_lingbot_video_pre_process_func is not None
    assert get_lingbot_video_post_process_func is not None
    assert _DIFFUSION_MODELS["LingBotVideoPipeline"] == (
        "lingbot_video",
        "pipeline_lingbot_video",
        "LingBotVideoPipeline",
    )
+    assert _DIFFUSION_PRE_PROCESS_FUNCS["LingBotVideoPipeline"] == "get_lingbot_video_pre_process_func"
    assert _DIFFUSION_POST_PROCESS_FUNCS["LingBotVideoPipeline"] == "get_lingbot_video_post_process_func"


@@ -58,6 +68,7 @@ def test_component_discovery_declarations():
    assert LingBotVideoPipeline._encoder_modules == ["text_encoder"]
    assert LingBotVideoPipeline._vae_modules == ["vae"]
    assert LingBotVideoPipeline.supports_step_execution is False
+    assert LingBotVideoPipeline.max_outputs_per_prompt == 1


 def test_extra_body_params_include_video_flow_shift_alias():
@@ -69,17 +80,47 @@ def test_extra_body_params_include_video_flow_shift_alias():
    assert "batch_cfg" in params


-def test_extract_prompt_string_and_mapping():
-    from vllm_omni.diffusion.models.lingbot_video.pipeline_lingbot_video import _extract_prompt
+def test_preprocess_preserves_geometry_and_converts_rgb(tmp_path):
+    from vllm_omni.diffusion.models.lingbot_video import (
+        get_lingbot_video_pre_process_func,
+    )

-    string_req = SimpleNamespace(prompt="a robot waves")
-    assert _extract_prompt(string_req) == ("a robot waves", None)
+    image_path = tmp_path / "reference.png"
+    Image.new("L", (48, 32), color=127).save(image_path)
+    request = OmniDiffusionRequest(
+        prompt={
+            "prompt": "a robot walks",
+            "modalities": ["video"],
+            "multi_modal_data": {"image": str(image_path)},
+        },
+        sampling_params=OmniDiffusionSamplingParams(),
+        request_id="lingbot-preprocess-test",
+    )

-    mapping_req = SimpleNamespace(prompt={"prompt": "a robot walks", "negative_prompt": "low quality"})
-    assert _extract_prompt(mapping_req) == ("a robot walks", "low quality")
+    result = get_lingbot_video_pre_process_func(SimpleNamespace())(request)

-    with pytest.raises(TypeError, match="Unsupported LingBot prompt type"):
-        _extract_prompt(SimpleNamespace(prompt=["not", "supported"]))
+    image = result.prompt["multi_modal_data"]["image"]
+    assert image.mode == "RGB"
+    assert image.size == (48, 32)
+
+
+def test_preprocess_marks_internal_image_warmup_as_video():
+    from vllm_omni.diffusion.models.lingbot_video import (
+        get_lingbot_video_pre_process_func,
+    )
+
+    request = OmniDiffusionRequest(
+        prompt={
+            "prompt": "dummy run",
+            "multi_modal_data": {"image": Image.new("RGB", (32, 32))},
+        },
+        sampling_params=OmniDiffusionSamplingParams(),
+        request_id=DUMMY_DIFFUSION_REQUEST_ID,
+    )
+
+    result = get_lingbot_video_pre_process_func(SimpleNamespace())(request)
+
+    assert result.prompt["modalities"] == ["video"]


 def test_check_inputs_accepts_t2v_shapes_and_rejects_invalid_values():
@@ -104,6 +145,59 @@ def test_postprocess_keeps_torch_by_default_and_converts_np_when_requested():
    array = post(frames, SimpleNamespace(output_type="np"))
    assert isinstance(array, np.ndarray)
    assert array.shape == (2, 4, 4, 3)
+    video = post({"video": frames}, SimpleNamespace(output_type="pt"))
+    assert set(video) == {"video"}
+    assert video["video"] is frames
+    image = post(
+        {"image": torch.ones(4, 4, 3)},
+        SimpleNamespace(output_type="pt"),
+    )
+    assert set(image) == {"image"}
+    assert isinstance(image["image"], np.ndarray)
+    assert image["image"].shape == (4, 4, 3)
+    latent = torch.ones(1, 4, 1, 2, 2)
+    latent_output = post({"image": latent}, SimpleNamespace(output_type="latent"))
+    assert set(latent_output) == {"image"}
+    assert latent_output["image"] is latent
+
+
+def test_postprocess_preserves_image_contract_for_serving():
+    from vllm_omni.diffusion.data import DiffusionOutput
+    from vllm_omni.diffusion.models.lingbot_video import get_lingbot_video_post_process_func
+    from vllm_omni.diffusion.output_formatter import (
+        format_diffusion_outputs,
+        normalize_diffusion_postprocess_output,
+    )
+    from vllm_omni.entrypoints.openai.api_server import _extract_images_from_result
+
+    sampling = OmniDiffusionSamplingParams(
+        num_inference_steps=1,
+        output_type="pt",
+        resolution=64,
+    )
+    request = OmniDiffusionRequest(
+        prompt={"prompt": "a robot", "modalities": ["image"]},
+        sampling_params=sampling,
+        request_id="lingbot-image-contract-test",
+    )
+    raw_output = {"image": torch.ones(8, 12, 3)}
+    processed = get_lingbot_video_post_process_func(SimpleNamespace())(raw_output, sampling)
+    normalized = normalize_diffusion_postprocess_output(processed)
+
+    assert normalized.primary_key == "image"
+    [result] = format_diffusion_outputs(
+        request=request,
+        od_config=SimpleNamespace(model_class_name="LingBotVideoPipeline"),
+        diffusion_output=DiffusionOutput(output=raw_output),
+        output_data=raw_output,
+        postprocess_output=normalized,
+    )
+
+    assert len(result.images) == 1
+    assert isinstance(result.images[0], np.ndarray)
+    endpoint_images = _extract_images_from_result(result)
+    assert len(endpoint_images) == 1
+    assert endpoint_images[0].size == (12, 8)


 def test_forward_resolves_t2v_sampling_and_flow_shift_alias():
@@ -116,7 +210,11 @@ def test_forward_resolves_t2v_sampling_and_flow_shift_alias():

    pipeline._generate = fake_generate
    req = _make_request_batch(
-        {"prompt": "a robot arm", "negative_prompt": "bad hands"},
+        {
+            "prompt": "a robot arm",
+            "negative_prompt": "bad hands",
+            "modalities": ["video"],
+        },
        height=192,
        width=320,
        num_frames=9,
@@ -129,10 +227,12 @@ def test_forward_resolves_t2v_sampling_and_flow_shift_alias():

    out = pipeline.forward(req)

-    assert torch.equal(out.output, torch.zeros(9, 192, 320, 3))
+    assert torch.equal(out.output["video"], torch.zeros(9, 192, 320, 3))
    assert len(calls) == 1
    call = calls[0]
    assert call["prompt"] == "a robot arm"
+    assert call["mode"].value == "t2v"
+    assert call["input_image"] is None
    assert call["negative_prompt"] == "bad hands"
    assert call["height"] == 192
    assert call["width"] == 320
@@ -194,6 +294,222 @@ def test_forward_prefers_shift_over_flow_shift_and_defaults_negative_prompt():
    assert calls[0]["negative_prompt"] == "default negative"


+@pytest.mark.parametrize(
+    ("prompt", "num_frames", "expected_mode", "output_key", "default_negative"),
+    [
+        (
+            {"prompt": "a still robot", "modalities": ["image"]},
+            1,
+            "t2i",
+            "image",
+            "default image negative",
+        ),
+        (
+            {
+                "prompt": "a moving robot",
+                "modalities": ["video"],
+                "multi_modal_data": {"image": Image.new("RGB", (48, 32), color="blue")},
+            },
+            9,
+            "ti2v",
+            "video",
+            "default negative",
+        ),
+    ],
+)
+def test_forward_dispatches_generation_mode_and_output(
+    prompt,
+    num_frames,
+    expected_mode,
+    output_key,
+    default_negative,
+):
+    pipeline = _make_pipeline()
+    calls = []
+
+    def fake_generate(**kwargs):
+        calls.append(kwargs)
+        return torch.zeros(192, 320, 3) if output_key == "image" else torch.zeros(num_frames, 192, 320, 3)
+
+    pipeline._generate = fake_generate
+    output = pipeline.forward(
+        _make_request_batch(
+            prompt,
+            height=192,
+            width=320,
+            num_frames=num_frames,
+            num_inference_steps=2,
+            guidance_scale=3.0,
+            seed=42,
+            output_type="pt",
+        )
+    )
+
+    assert set(output.output) == {output_key}
+    assert calls[0]["mode"].value == expected_mode
+    assert calls[0]["negative_prompt"] == default_negative
+    if expected_mode == "ti2v":
+        assert isinstance(calls[0]["input_image"], Image.Image)
+
+
+def test_batched_cfg_pads_image_conditioned_embeddings_and_masks():
+    from vllm_omni.diffusion.models.lingbot_video.pipeline_lingbot_video import (
+        _batch_cfg_prompt_inputs,
+    )
+
+    positive = torch.ones(1, 7, 4)
+    positive_mask = torch.ones(1, 7, dtype=torch.long)
+    negative = torch.full((1, 9, 4), 2.0)
+    negative_mask = torch.ones(1, 9, dtype=torch.long)
+
+    embeds, mask = _batch_cfg_prompt_inputs(
+        positive,
+        positive_mask,
+        negative,
+        negative_mask,
+    )
+
+    assert embeds.shape == (2, 9, 4)
+    assert mask.shape == (2, 9)
+    assert torch.equal(mask[0], torch.tensor([1, 1, 1, 1, 1, 1, 1, 0, 0]))
+    assert torch.equal(mask[1], torch.ones(9, dtype=torch.long))
+
+
+class _RecordingTransformer(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.anchor = nn.Parameter(torch.zeros(()))
+        self.config = SimpleNamespace(in_channels=1)
+        self.latent_inputs = []
+
+    def forward(
+        self,
+        latents,
+        timestep,
+        prompt_embeds,
+        *,
+        encoder_attention_mask,
+        return_dict,
+    ):
+        del timestep, prompt_embeds, encoder_attention_mask, return_dict
+        self.latent_inputs.append(latents.detach().clone())
+        return (torch.zeros_like(latents),)
+
+
+class _CorruptingScheduler:
+    sigma_max = 1.0
+    sigma_min = 0.0
+
+    def set_timesteps(self, num_inference_steps, *, device, shift, **kwargs):
+        del num_inference_steps, shift, kwargs
+        self.timesteps = torch.tensor([1000.0, 500.0], device=device)
+
+    def step(
+        self,
+        noise_pred,
+        timestep,
+        latents,
+        *,
+        return_dict,
+        generator,
+    ):
+        del noise_pred, timestep, return_dict, generator
+        return (latents + 7.0,)
+
+
+@pytest.mark.parametrize(
+    ("guidance_scale", "batch_cfg", "expected_prompts", "expected_transformer_calls"),
+    [
+        pytest.param(2.0, False, ["positive", "negative"], 4, id="sequential-cfg"),
+        pytest.param(2.0, True, ["positive", "negative"], 2, id="batched-cfg"),
+        pytest.param(1.0, False, ["positive"], 2, id="cfg-disabled"),
+    ],
+)
+def test_ti2v_reinjects_clean_prefix_across_cfg_modes(
+    guidance_scale,
+    batch_cfg,
+    expected_prompts,
+    expected_transformer_calls,
+):
+    from vllm_omni.diffusion.models.lingbot_video import (
+        LingBotGenerationMode,
+        LingBotImageCondition,
+    )
+
+    pipeline = _make_pipeline()
+    pipeline.transformer = _RecordingTransformer()
+    pipeline.scheduler = _CorruptingScheduler()
+    pipeline.progress_bar = lambda timesteps: timesteps
+    pipeline.prepare_latents = lambda *args, **kwargs: torch.zeros(1, 1, 3, 1, 1)
+    vlm_image = Image.new("RGB", (32, 32), color="red")
+    condition = LingBotImageCondition(
+        vlm_image=vlm_image,
+        clean_latent=torch.full((1, 1, 1, 1, 1), 5.0),
+    )
+    pipeline.prepare_ti2v_image_condition = lambda *args, **kwargs: condition
+    encode_calls = []
+
+    def fake_encode(prompt, *, images, device):
+        del device
+        encode_calls.append((prompt, images))
+        return torch.ones(1, 2, 4), torch.ones(1, 2, dtype=torch.long)
+
+    pipeline.encode_prompt = fake_encode
+
+    latents = pipeline._generate(
+        prompt="positive",
+        mode=LingBotGenerationMode.TI2V,
+        input_image=vlm_image,
+        negative_prompt="negative",
+        height=16,
+        width=16,
+        num_frames=9,
+        num_inference_steps=2,
+        guidance_scale=guidance_scale,
+        shift=3.0,
+        output_type="latent",
+        batch_cfg=batch_cfg,
+    )
+
+    assert [call[0] for call in encode_calls] == expected_prompts
+    assert all(call[1] == [vlm_image] for call in encode_calls)
+    assert len(pipeline.transformer.latent_inputs) == expected_transformer_calls
+    for item in pipeline.transformer.latent_inputs:
+        expected_prefix = condition.clean_latent.expand(item.shape[0], -1, -1, -1, -1)
+        assert torch.equal(item[:, :, :1], expected_prefix)
+    assert torch.equal(latents[:, :, :1], condition.clean_latent)
+    assert torch.equal(latents[:, :, 1:], torch.full((1, 1, 2, 1, 1), 14.0))
+
+
+def test_t2i_decodes_and_returns_the_unique_frame():
+    from vllm_omni.diffusion.models.lingbot_video import LingBotGenerationMode
+
+    pipeline = _make_pipeline()
+    pipeline.transformer = _RecordingTransformer()
+    pipeline.scheduler = _CorruptingScheduler()
+    pipeline.progress_bar = lambda timesteps: timesteps
+    pipeline.prepare_latents = lambda *args, **kwargs: torch.zeros(1, 1, 1, 2, 2)
+    pipeline.encode_prompt = lambda *args, **kwargs: (
+        torch.ones(1, 2, 4),
+        torch.ones(1, 2, dtype=torch.long),
+    )
+    pipeline._decode_latents = lambda latents: torch.ones(1, 16, 16, 3)
+
+    image = pipeline._generate(
+        prompt="a still robot",
+        mode=LingBotGenerationMode.T2I,
+        height=16,
+        width=16,
+        num_frames=1,
+        num_inference_steps=2,
+        guidance_scale=1.0,
+        shift=3.0,
+        output_type="pt",
+    )
+
+    assert image.shape == (16, 16, 3)
+
+
 def test_forward_rejects_multi_request_batches():
    pipeline = _make_pipeline()
    first = OmniDiffusionRequest(
@@ -211,6 +527,45 @@ def test_forward_rejects_multi_request_batches():
        pipeline.forward(DiffusionRequestBatch([first, second]))


+def test_forward_marks_invalid_request_contract_as_client_error():
+    from vllm_omni.errors import OmniClientError
+
+    pipeline = _make_pipeline()
+    request = _make_request_batch(
+        {
+            "prompt": "unsupported image edit",
+            "modalities": ["image"],
+            "multi_modal_data": {"image": Image.new("RGB", (32, 32), color="blue")},
+        },
+        height=192,
+        width=320,
+        num_frames=1,
+    )
+
+    with pytest.raises(OmniClientError, match="does not accept") as exc_info:
+        pipeline.forward(request)
+
+    assert exc_info.value.status_code == 400
+
+
+def test_forward_marks_unsupported_output_count_as_client_error():
+    from vllm_omni.errors import OmniClientError
+
+    pipeline = _make_pipeline()
+    request = _make_request_batch(
+        {"prompt": "motion", "modalities": ["video"]},
+        height=192,
+        width=320,
+        num_frames=9,
+        num_outputs_per_prompt=2,
+    )
+
+    with pytest.raises(OmniClientError, match="one output per prompt") as exc_info:
+        pipeline.forward(request)
+
+    assert exc_info.value.status_code == 400
+
+
 def test_load_weights_rejects_external_weight_stream():
    pipeline = _make_pipeline()
Author	SHA1	Message	Date
Zhou TaichangandClaude Opus 5	2ce15ab7a0	[Rebase] Rebase to vllm 0.27.0 (#5976 ) Signed-off-by: tzhouam <tzhouam@connect.ust.hk> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>	2026-08-11 15:23:46 +00:00
Nick Cao	584d78c673	[Test] Add tiny model builder for FluxKontextPipeline (#5823 ) Signed-off-by: Nick Cao <ncao@redhat.com>	2026-08-11 10:59:09 -04:00
Jerry Song	89015d5cad	[Feature]: Adds distilled LoRA support for diffusion models (#2783 ) Signed-off-by: Songrui625 <songrui625@gmail.com>	2026-08-11 14:23:43 +00:00
amy-why-3459	348cf95a89	[CI][MiniCPM-o] Align MiniCPM-o 4.5 online serving tests with minicpm… (#6056 ) Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>	2026-08-11 21:11:39 +08:00
R0CKSTAR	22dd96ec5c	[Hardware][MUSA] Complete Omni platform interfaces (#6058 ) Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>	2026-08-11 12:46:42 +00:00
R0CKSTAR Hongsheng Liu amy-why-3459	e04210d690	[Hardware][MUSA] Enable Qwen3-Omni ModelOpt FP8 inference (#5671 ) Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com> Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com> Co-authored-by: amy-why-3459 <wuhaiyan17@huawei.com>	2026-08-11 17:19:09 +08:00
Kristoffer Didan Deng Hongsheng Liu	4ab1b4ff9c	[Perf][Flux2][HunyuanVideo1.5] Skip attention-mask to avoid varlen path (#4645 ) Signed-off-by: kTorp <Kristoffer.Torp@amd.com> Signed-off-by: Didan Deng <33117903+wtomin@users.noreply.github.com> Co-authored-by: Didan Deng <33117903+wtomin@users.noreply.github.com> Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-11 15:31:02 +08:00
Weiyan Didan Deng Hongsheng Liu	06e733d3cd	[Model] Add LongCat-Video-Avatar-1.5 ai2v, at2v support (#4099 ) Signed-off-by: weiyan <weiyanlin117@gmail.com> Signed-off-by: Didan Deng <33117903+wtomin@users.noreply.github.com> Co-authored-by: Didan Deng <33117903+wtomin@users.noreply.github.com> Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-11 15:17:17 +08:00
Wang fuyinandHongsheng Liu	af3faa6c1e	[Diffusion] Add Worker contracts and RPC plumbing for Scheduler-managed paged KV cache (#5550 ) Signed-off-by: Acerak01-fy <wfy2003324@163.com> Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com> Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-11 14:55:42 +08:00
Ruirui Yang \| Rein psv666 Cursor amy-why-3459	322600bfb2	[CI][MiniCPM-o] Add MiniCPM-o 4.5 accuracy and performance coverage (#5524 ) Signed-off-by: R2-Y <ruiruyang2@gmail.com> Signed-off-by: psv666 <2693925048@qq.com> Co-authored-by: psv666 <2693925048@qq.com> Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: amy-why-3459 <wuhaiyan17@huawei.com>	2026-08-11 14:53:49 +08:00
Hongsheng Liu	c4a14091c9	[Doc] Explain repository skills for agentic contributions (#6029 ) Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-11 14:18:41 +08:00
yiminghub2024	6d4e4ffc7c	Add MiniMax-H3 recipe for RTX 4090 setup (#5850 ) Signed-off-by: yiminghub2024 <482890@qq.com>	2026-08-11 14:08:51 +08:00
Zhou248	f97af75d1d	[Model] Batch MiniCPM-o Talker codec sampling (#5792 ) Signed-off-by: Zhou248 <62542847+Zhou248@users.noreply.github.com>	2026-08-11 14:05:02 +08:00
Hongsheng Liu	73a312d5b6	[Bugfix] Normalize NumPy image outputs before saving (#6031 ) Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-11 13:48:03 +08:00
Zijun HuangandWeiQing Chen	1a56953d7a	[NPU][Model] Add distilled 4-step sigma schedule support for MiniMax H3 t2va (#5991 ) Signed-off-by: Zijun Huang <zijunhua@usc.edu> Co-authored-by: WeiQing Chen <40507679+david6666666@users.noreply.github.com>	2026-08-11 13:46:10 +08:00
WeiQing Chen	1f65814fab	[CI] Align MiniMax H3 Ref2VA inputs and add I2VA/Ref2VA accuracy coverage (#5978 ) Signed-off-by: david6666666 <530634352@qq.com>	2026-08-11 13:36:37 +08:00
boatmanandYueqian Lin	4f9b65074d	[Perf][Qwen3-TTS]Cached Incremental Decode (#5202 ) Signed-off-by: boatman <109857087+sphinxkkkbc@users.noreply.github.com> Co-authored-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>	2026-08-11 05:24:04 +00:00
Juan Pablo ZuluagaandYueqian Lin	12a5f6fbf3	[Perf][TTS] step_audio2: keep streaming tokens on-device and batch audio-feature length syncs (#5067 ) Signed-off-by: JuanPZuluaga <juanz9312@gmail.com> Co-authored-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>	2026-08-10 21:24:25 -07:00
Juan Pablo ZuluagaandYueqian Lin	53afbf3a0c	[Perf][TTS] voxtral_tts hot path: cudagraph opt-out, host-sync removal, kernel caches (#5175 ) Signed-off-by: JuanPZuluaga <juanz9312@gmail.com> Co-authored-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>	2026-08-10 21:03:32 -07:00
Shenglei Fu Claude Opus 4.8 Yueqian Lin	ee118ab492	[Bugfix] Per-replica fault isolation: keep API server alive on single-stage death (#4285 ) (#4583 ) Signed-off-by: Shenglei Fu <sfu@confluent.io> Signed-off-by: Shenglei Fu <117230642+ShengleiFu@users.noreply.github.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>	2026-08-10 23:00:18 -04:00
fan2956andFan	c27623c29c	[Bugfix] rainfusion support all shape video (#6000 ) Signed-off-by: Fan <fan@FandeMacBook-Air.local> Co-authored-by: Fan <fan@FandeMacBook-Air.local>	2026-08-11 05:10:50 +08:00
Yueqian Lin	f8fa86fb1f	[BugFix] Don't fail the TTS ratchet when the branch count goes down (#6008 ) Signed-off-by: Yueqian Lin <linyueqian@outlook.com>	2026-08-11 05:08:31 +08:00
ooooooyeandHongsheng Liu	422d8fd8f9	[Bugfix][Diffusion] Restore supports_packed_mask_free on teacache FakeBackend (#5997 ) Signed-off-by: brandneway <gyuan4892@gmail.com> Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-10 22:20:37 +08:00
Hongsheng Liu	62e239bf1e	[Docs] Fix broken attention backend link (#5998 ) Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-10 21:55:01 +08:00
ooooooye	740f45abcd	[Perf][Diffusion] MiniMax-H3: opt into packed varlen attention on NPU to eliminate quadratic mask materialization (#5891 ) Signed-off-by: brandneway <gyuan4892@gmail.com>	2026-08-10 19:47:03 +08:00
Hongsheng Liu hsliu_ustc Copilot Autofix powered by AI	5e4b9f8c6e	[Misc] Align CODEOWNERS with module/feature design docs (#5958 ) Signed-off-by: hsliu_ustc <hsliu_ustc@noreply.gitcode.com> Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com> Co-authored-by: hsliu_ustc <hsliu_ustc@noreply.gitcode.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>	2026-08-10 19:25:03 +08:00
Hongsheng Liu	ae811043e9	docs: keep shared task examples in navigation (#5987 ) Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-10 19:13:52 +08:00
R0CKSTAR	636d44e272	[Perf][MUSA] Restore MiniMax-H3 dynamic RoPE fusion (#5881 ) Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>	2026-08-10 17:24:46 +08:00
yiminghub2024	8e44e78038	Add Ref2VA measurements to the MiniMax-H3 DGX Spark (GB10) recipe (#5972 ) Signed-off-by: yiminghub2024 <482890@qq.com>	2026-08-10 15:51:15 +08:00
Hongsheng Liu	b98b7c8529	[Docs] Preserve generated quantization URLs and link recipes in supported models (#5969 ) Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-10 15:45:50 +08:00
bjf-frzandAlicia	1ee2fff532	Update Wan2.2 I2V performance baselines (#5977 ) Signed-off-by: bjf-frz <frz123db@gmail.com> Co-authored-by: Alicia <115451386+congw729@users.noreply.github.com>	2026-08-10 15:24:59 +08:00
Zeyu Huang \| 黃澤宇andHongsheng Liu	33dd9a5a3e	[Core][Refactor][Diffusion] refactor request scheduler's admission wait policy (#5843 ) Signed-off-by: Huang, Zeyu <11222265+fhfuih@users.noreply.github.com> Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-10 15:06:27 +08:00
Nguyen Kha Nhat Long Samit Gao Han Hongsheng Liu	c588208cc6	[bugfix] Support lora request for non diffusion model (#5374 ) Signed-off-by: knlnguyen1802 <knlnguyen1802@gmail.com> Co-authored-by: Samit <285365963@qq.com> Co-authored-by: Gao Han <hgaoaf@connect.ust.hk> Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-10 11:27:51 +08:00
wangyuandamy-why-3459	e28bf8ff8b	[CI] Re-enable bagel shared-memory connector test (#5475 ) (#5898 ) Signed-off-by: wangyu <410167048@qq.com> Co-authored-by: amy-why-3459 <wuhaiyan17@huawei.com>	2026-08-10 11:26:10 +08:00
WeiQing Chen	e1aa6eae75	docs: update vLLM-Omni WeChat QR code (#5959 ) Signed-off-by: David Chen <530634352@qq.com>	2026-08-10 10:04:24 +08:00
wtz2333andClaude	36135ef09a	[diffusion][feature] Add LingBot-Video T2I and TI2V generation modes (#5311 ) Signed-off-by: wtz2333 <2955110911@qq.com> Co-authored-by: Claude <noreply@anthropic.com>	2026-08-10 10:02:31 +08:00
boatman	11f633aa83	[Refactor] Add Ming Flash Omni TTS adapter (#5746 ) Signed-off-by: boatman <109857087+sphinxkkkbc@users.noreply.github.com>	2026-08-10 01:47:09 +00:00
Yukim1 zwhzzz0821 Hongsheng Liu	6ae7ff7846	[Diffusion] Prepare HunyuanImage3 for Scheduler-managed paged KV cache (#5541 ) Signed-off-by: zwhzzz0821 <zwhzzz0821@users.noreply.github.com> Co-authored-by: zwhzzz0821 <zwhzzz0821@users.noreply.github.com> Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>	2026-08-10 08:48:56 +08:00
yiminghub2024	615de4e041	[doc]Add recipe for MiniMax-H3 on RTX PRO 6000 (#5863 ) Signed-off-by: yiminghub2024 <482890@qq.com>	2026-08-09 23:52:40 +08:00
yiminghub2024	be48b0797e	Add MiniMax-H3 recipe for DGX Spark (GB10) (#5946 ) Signed-off-by: yiminghub2024 <482890@qq.com>	2026-08-09 23:15:05 +08:00
Diya Pengandlishunyang12	45629a8e26	[MiniMax-H3] TeaCache support and Cache-DiT validation (#5840 ) Signed-off-by: dpeng123 <diyapeng123@gmail.com> Signed-off-by: lishunyang12 <lishunyang12@163.com> Co-authored-by: lishunyang12 <lishunyang12@163.com>	2026-08-09 15:07:21 +00:00