anodyine/vllm-omni
1
0 Fork 0
Code Issues 4 Pull Requests 2 Actions Packages Projects Releases Wiki Activity
3,228 Commits 235 Branches 31 Tags
44ea27c8094095bbffd88fa3befdfaa55ba4bc50
Commit Graph
3228 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Isotr0py 44ea27c809 fix pre-commit 
Signed-off-by: Isotr0py <Isotr0py@outlook.com>
 2026-09-23 06:22:17 +00:00
Isotr0py 0c82bb1315 Gate persistent DiT staging on decode-graph models 
ModelLevelOffloadBackend passed persistent_dit_staging=(device is CUDA) for
every model, so Qwen-Image 2.0/FLUX/WAN-style sequential offload kept DiT
device storage alive across swaps instead of freeing VRAM. Only retain
staging when a DiT actually captures weight pointers (enable_cuda_graph_decode);
other models keep plain move semantics. Add tests pinning both behaviors.
 2026-09-22 08:58:32 +00:00
Isotr0py be8a72f0a6 Stack torch.compile with CUDA graph decode instead of choosing one 
Regionally compile the DiT blocks even when enable_cuda_graph_decode is set
and let graph capture record the compiled (fused) kernels; inductor's own
cudagraphs stay off so the two graph layers never nest. Measured on GB200
(1024x1024, 50 steps, seed 42, BF16): combo 3.04s vs compile-only 3.2s,
graph-only 4.0s, eager 4.3s per image; 44.5-45.6 dB PSNR vs eager.
 2026-09-22 08:58:32 +00:00
Isotr0py 45aa5b16d9 Rename internal temperal_upsample to temporal_upsample in Qwen-Image-2.1 VAE 
Match the Qwen-Image 2.0 split: decoder kwargs/attrs use the correct
temporal_upsample spelling while the checkpoint-facing temperal_downsample
config key is kept as-is. Fixes pre-commit typos rewriting the name.
 2026-09-22 08:58:31 +00:00
Isotr0py 8e0695f362 Merge branch 'main' into qwen-image-2.1 
Signed-off-by: Isotr0py <Isotr0py@outlook.com>
 2026-09-22 10:11:12 +08:00
Hoseung KimandHoseung Kim d4570afabf [Frontend] Reject built-in name collisions in voice upload and add VLLM_OMNI_SPEAKER_REGISTRATION_POLICY (#6848) 
Signed-off-by: Hoseung Kim <hoseung.kim@navercorp.com>
Co-authored-by: Hoseung Kim <hoseung.kim@navercorp.com>
 2026-09-21 18:35:07 -07:00
Tianyao WuClaudeAlicia
 		085cf0ff34 [Bugfix] Catch asyncio.TimeoutError so duplex serving starts on Python 3.10 (#7412) 
Signed-off-by: Tianyao Wu <rayroy31@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Alicia <115451386+congw729@users.noreply.github.com>
 2026-09-21 18:34:58 -07:00
Nick CaoandCodex 1d3dc8bf83 [Misc] Refactor static stage metadata lookups (#7602) 
Signed-off-by: Nick Cao <ncao@redhat.com>
Co-authored-by: Codex <noreply@openai.com>
 2026-09-21 18:34:50 -07:00
andyluo7 1586c41ae4 [CI][ROCm] Shorten the blocking CosyVoice test path (#7748) 
Signed-off-by: andyluo7 <andy.luo@amd.com>
 2026-09-21 18:34:40 -07:00
Armaan Amatya cb31e0774f [Misc] Deprecate VLLM_OMNI_VOXCPM_CODE_PATH (#7779) 
Signed-off-by: Armaan Amatya <armaanamatya2014@gmail.com>
 2026-09-21 18:34:31 -07:00
0z5aand0z5a 8bc0c38b90 [Bugfix] Add opt-in MiniCPM-o 4.5 Stage-0 sliding window (#7631) 
Signed-off-by: 0z5a <dezhen.lu@student.uni-tuebingen.de>
Signed-off-by: 0z5a <dezhen.lu@uni-tuebingen.de>
Co-authored-by: 0z5a <dezhen.lu@uni-tuebingen.de>
 2026-09-21 18:34:22 -07:00
shihongzhi c11596c637 [Model] MammothModa2: enable Cache-DiT acceleration for the DiT stage (#7291) 
Signed-off-by: shihongzhi <shi65881583@gmail.com>
 2026-09-21 18:34:12 -07:00
Isotr0py c05a3527f5 ooops 
Signed-off-by: Isotr0py <Isotr0py@outlook.com>
 2026-09-22 01:26:57 +00:00
Isotr0py 3e274a0024 Merge remote-tracking branch 'origin/main' into qwen-image-2.1 
# Conflicts:
#	tests/entrypoints/openai_api/test_image_server.py
#	vllm_omni/entrypoints/openai/api_server.py
 2026-09-22 01:23:45 +00:00
jingchengtian 0b5f8832b9 [tts][feature][NPU] Add 2-NPU and 3-NPU disaggregated deploy profiles for MOSS-TTS-Local (#7052) 
Signed-off-by: jingchengtian <tjc1995@126.com>
 2026-09-21 23:14:52 +00:00
Sun 381b12fa87 [Bugfix] Keep --no-guardrails out of diffusion stage config (#7917) 
Signed-off-by: levius <2114377220@qq.com>
 2026-09-21 19:18:30 +00:00
Sy03 473da2cf82 [Perf][Qwen3-TTS] Enable event-driven orchestration by default (#7088) 
Signed-off-by: Sy03 <1370724210@qq.com>
 2026-09-21 12:09:19 -07:00
Sy03 60a933b9ab [Model] Default Qwen3-TTS to experimental Model Runner V2 (#7930) 
Signed-off-by: Sy03 <1370724210@qq.com>
 2026-09-21 18:56:06 +00:00
vOv d8d162d810 [Core][Diffusion] Add pause_generation(mode="keep") for diffusion stages (#7685) 
Signed-off-by: cr-gao <gaochenrui@sjtu.edu.cn>
 2026-09-22 01:31:22 +08:00
NumberWan 62b15142bd [Model] Integrate AURA into Full-Duplex Runtime (#7633) 2026-09-21 09:26:54 -07:00
wuli666 d397a2f547 [Bugfix] Refresh duplex client defaults after session updates (#7784) 
Signed-off-by: wuli666 <421774554@qq.com>
 2026-09-21 16:24:42 +00:00
wangyu 279b1b9687 [Bugfix] Isolate weekly CPU tests from inherited attention env and hf_api mocks (#7565) 
Signed-off-by: wangyu <410167048@qq.com>
Signed-off-by: [Your Name] <your.email@example.com>
Signed-off-by: [Your Name] [Your Email]
 2026-09-21 15:55:00 +00:00
0z5a 2c2cffd2aa [Bugfix] Preserve NextStep rectangular latent dimensions (#7853) 
Signed-off-by: 0z5a <dezhen.lu@student.uni-tuebingen.de>
 2026-09-21 08:26:14 -07:00
akshatvishu e01655f4a9 [BugFix] Fix diffusion TTS voice error responses (#7798) 
Signed-off-by: akshatvishu <akshatnayak197@gmail.com>
 2026-09-21 15:09:32 +00:00
wangyu ef644b6112 [CI/Build] Promote LingBot-Video dense smoke to ready/merge and wire L4 expansion (#7892) 
Signed-off-by: wangyu <410167048@qq.com>
 2026-09-21 07:45:32 -07:00
WeiQing Chen 44ea368843 [Doc] Update vLLM-Omni WeChat QR code (#7911) 
Signed-off-by: David Chen <530634352@qq.com>
 2026-09-21 14:18:28 +00:00
Zhou TaichangandClaude Opus 5 856d19f42c [Benchmark] Add LingBot-World realtime streaming benchmark (#7645) 
Signed-off-by: tzhouam <tzhouam@connect.ust.hk>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
 2026-09-21 13:12:00 +00:00
Weiming Liaochen hongweichen hongwei
 		fa1d03ab8d [Frontend] Add MiniMax-H3 latent editing workflows (WF-05) (#7898) 
Signed-off-by: chen hongwei <1792043268@qq.com>
Signed-off-by: Weiming Liao <liaowm5@gmail.com>
Co-authored-by: chen hongwei <1792043268@qq.com>
Co-authored-by: chen hongwei <54873389+avicii-forever@users.noreply.github.com>
 2026-09-21 13:07:21 +00:00
tsinghua-code e5c0afc994 [Model] Enable online FP8 linears for LingBot World (#7549) 
Signed-off-by: tsinghua-code <zr-liu25@mails.tsinghua.edu.cn>
 2026-09-21 12:21:46 +00:00
zhumingjue138 5e28163ba5 [Test] add stability test case for High-priority model (#7571) 
Signed-off-by: zhumingjue <zhumingjue@huawei.com>
 2026-09-21 04:03:18 -07:00
Zhou TaichangandClaude Fable 5.1 5356357cf0 [LingBot World] Shard the streaming VAE decode across the Ulysses ranks (#7651) 
Signed-off-by: tzhouam <tzhouam@connect.ust.hk>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
 2026-09-21 19:00:46 +08:00
wtz2333 8efd3c2f6d [Feature] Add experimental LingBot World last-step KV reuse (#7816) 
Signed-off-by: wtz2333 <2955110911@qq.com>
 2026-09-21 17:02:15 +08:00
shiyichuanandmershi fb2e3535af [Bugfix][Engine] Preserve text_encoder_tp_size in OmniEngineArgs (#7652) 
Signed-off-by: mershi <mershi@tencent.com>
Co-authored-by: mershi <mershi@tencent.com>
 2026-09-21 17:01:03 +08:00
Isotr0py 3d6a1c4d9a Address review: examples policy, SPDX headers, recipe docs, test coverage 
- Register QwenImage21Pipeline in model_extras _EXTRA_SPECS with an
  image-to-image prompt builder that forwards height/width
- Rename run_qwen_image_21_edit.sh to run_qwen_image_21.sh
- Fix SPDX-FileCopyrightText headers
- Recipe: document mixed-KV-phase admission, add head-vs-diffusers A/B
  comparison section (measurements pending)
- Add test_registry.py and test_pipeline.py (preprocess, prepare_latents
  ordering, mixed-phase _assemble_kv_cache, post-process); extend
  test_batch_compatibility.py
 2026-09-21 08:36:42 +00:00
Isotr0py d3cf26c1b1 Wire enable_cuda_graph_decode through engine args, stage config and CLI 
Follows the vae_use_tiling path: OmniDiffusionConfig field, OmniEngineArgs,
StageDeployConfig, _DiffusionConfigProjection and --[no-]enable-cuda-graph-decode.
Documents interplay with enforce_eager and cudagraph_mode.
 2026-09-21 08:36:34 +00:00
Isotr0py a04b95bf6b Address review: tuple image input, NPU index_select rope, per-key decode graph ownership 
- Normalize tuple image inputs in pre_process_func (empty tuple -> T2I path)
- Use torch.index_select in QwenImage21Rope for NPU compatibility (#6320 style)
- Refuse concurrent aliasing of shared decode-graph prefix KV entries: track a
  single live owner per key, non-owners fall back to eager decode
 2026-09-21 08:36:28 +00:00
WeiQing Chen 84977d9543 [Bugfix][MiniMax-H3] Fix Hopper keyframe and modulation precision (#7913) 
Signed-off-by: david6666666 <530634352@qq.com>
 2026-09-21 16:26:09 +08:00
Isotr0py 9174554e40 Merge remote-tracking branch 'origin/main' into qwen-image-2.1 
# Conflicts:
#	vllm_omni/diffusion/executor/multiproc_executor.py
#	vllm_omni/diffusion/offloader/sequential_backend.py
 2026-09-21 07:32:04 +00:00
Ding ZuhaoandCursor 154a1f25f5 [Bugfix] Honor the requested output canvas in BAGEL img2img (#7287) 
Signed-off-by: Ding Zuhao <e1583181@u.nus.edu>
Co-authored-by: Cursor <cursoragent@cursor.com>
 2026-09-21 07:29:24 +00:00
Zhou TaichangandClaude Fable 5.1 86fcfe95f1 [LingBot World] Eleven bit-identical removals of repeated work in the realtime path (#7648) 
Signed-off-by: tzhouam <tzhouam@connect.ust.hk>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
 2026-09-21 06:12:45 +00:00
Sy03 a9e2117f5e [Core][Model] Enable Qwen3-TTS MRV2 and optimize the TTS pipeline (#7781) 
Signed-off-by: Sy03 <1370724210@qq.com>
 2026-09-20 23:03:08 -07:00
DMAN d139c600c7 [Model] Add online FP8 support for Boogu-Image MLLM (#7342) 
Signed-off-by: Dmaner <2663515256@qq.com>
 2026-09-21 05:13:37 +00:00
Abhinav Raichur 1b87115c9e feat(comfyui): add MiniMax H3 first and last frame inputs (#7449) 
Signed-off-by: avraichur96 <avraichur96@gmail.com>
Signed-off-by: Abhinav Raichur  <avraichur96@gmail.com>
 2026-09-21 10:24:25 +08:00
Wallbreazzzamy-why-3459Gao Han
 		ae3880f0e3 [Performance] Capture MOSS-TTS codec streaming decode with NPUGraph (#7280) 
Signed-off-by: Wallbreazzz <110282866+Wallbreazzz@users.noreply.github.com>
Co-authored-by: amy-why-3459 <wuhaiyan17@huawei.com>
Co-authored-by: Gao Han <hgaoaf@connect.ust.hk>
 2026-09-21 06:38:23 +08:00
Zeyu Huang | 黃澤宇 e36babd48f [Core][Diffusion] Camera interaction for diffusion streaming generation (LingBot World 2 as example) (#7198) 
Signed-off-by: Huang, Zeyu <11222265+fhfuih@users.noreply.github.com>
 2026-09-20 16:19:41 +00:00
汪志鹏 139a47a578 [Model] Add MiniMax-H3 long-video latent continuation with driving audio (#7838) 
Signed-off-by: princepride <wangzhipeng628@gmail.com>
 2026-09-20 23:44:33 +08:00
DanaerLeeandzijianc2 16734bbdf3 [Bugfix] Route text-only chat as per-request comprehension in HunyuanImage3 AR sampler (#6111) 
Signed-off-by: MrlixiangWE <mrdanaer@gmail.com>
Co-authored-by: zijianc2 <157244773+zijianc2@users.noreply.github.com>
 2026-09-20 15:23:53 +00:00
Wu JIAZHENAcerak01-fyamy-why-3459
 		102ba716bf [Feat] Add native Mooncake KV transfer from AR to DiT (#7166) 
Signed-off-by: asukaqaq-s <1311722138@qq.com>
Signed-off-by: Acerak01-fy <wfy2003324@163.com>
Co-authored-by: Acerak01-fy <wfy2003324@163.com>
Co-authored-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-09-20 14:44:55 +00:00
psv666 2d0b03f717 [Bugfix] Fix Qwen3-Omni realtime playback interruption and conversation history (#7791) 
Signed-off-by: psv666 <2693925048@qq.com>
 2026-09-20 14:42:17 +00:00
RyanYun09 d3396fc5fb # [Bugfix] Replace ffmpeg subprocess with PyAV to avoid HEVC decoder hang (#7364) (#7504) 
Signed-off-by: RyanYun09 <318555231+RyanYun09@users.noreply.github.com>
 2026-09-20 14:37:09 +00:00