anodyine/vllm-omni
1
0 Fork 0
Code Issues 4 Pull Requests 2 Actions Packages Projects Releases Wiki Activity
1,419 Commits 235 Branches 31 Tags
74522ca3ece4cacc50b9935d9da0c08d24848975
Commit Graph
1419 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Yueqian Lin 74522ca3ec Merge branch 'main' into feat/moss-tts-nano 
Signed-off-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
 2026-04-24 11:00:13 -04:00
Sy03shimei.jmhSYLAR
 		7b876324c2 [Feature] Streaming video input with EVS frame filtering (RFC #2201 Phase 2-4) (#2342) 
Signed-off-by: Sy03 <1370724210@qq.com>
Co-authored-by: shimei.jmh <shimei.jmh@taobao.com>
Co-authored-by: SYLAR <125541396+lishunyang12@users.noreply.github.com>
 2026-04-24 22:54:35 +08:00
amy-why-3459 b8bd6fbbdb [Refactor] Remove redundant benchmarks from Qwen3-Omni (#3108) 
Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-04-24 21:22:03 +08:00
汪志鹏andSYLAR 37e5954acf [Config Refactor]: Remove bagel yaml (#2936) 
Signed-off-by: princepride <wangzhipeng628@gmail.com>
Signed-off-by: 汪志鹏 <wangzhipeng628@gmail.com>
Co-authored-by: SYLAR <125541396+lishunyang12@users.noreply.github.com>
 2026-04-24 21:21:30 +08:00
Hacoandlishunyang 22ffb85063 [Config Refactor] sentinel default precedence (#3078) 
Signed-off-by: lishunyang <lishunyang12@163.com>
Signed-off-by: xiaohajiayou <923390377@qq.com>
Co-authored-by: lishunyang <lishunyang12@163.com>
 2026-04-24 20:07:24 +08:00
Lancer b33d536f63 [Bugfix] graceful shutdown for multi-stage engine processes (#3001) 
Signed-off-by: Lancer <maruixiang6688@gmail.com>
 2026-04-24 20:01:01 +08:00
Lancer 0fe76857ac [Bugfix] fix GLM-Image multi-stage online generation fail (#3084) 
Signed-off-by: Lancer <maruixiang6688@gmail.com>
 2026-04-24 20:00:14 +08:00
Yueqian Lin e8d22d4fa8 docs: update add-tts-model skill and contributing guide with single-stage patterns (#2806) 
Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
Signed-off-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
 2026-04-24 19:56:32 +08:00
Alex Brooks 77e8ba0a7b [Bugfix] Fix Flux2klein Text Input Processing (#3098) 
Signed-off-by: Alex Brooks <albrooks@redhat.com>
 2026-04-24 19:54:41 +08:00
Lancer c6edb34d04 [Bugfix] T5 text encoder to render correct text in FLUX.1-dev (#2760) 
Signed-off-by: Lancer <maruixiang6688@gmail.com>
 2026-04-24 18:01:52 +08:00
bjf-frz fb4204ebf6 [CI][Perf] Add Wan22 i2v perf nightly ci (#3063) 
Signed-off-by: bjf-frz <frz123db@gmail.com>
 2026-04-24 16:58:45 +08:00
bjf-frz e24f86ae2e [Enhancement]modify profiling.md (#3051) 
Signed-off-by: bjf-frz <frz123db@gmail.com>
 2026-04-24 16:30:24 +08:00
NumberWan 678af19240 [Feature] Coordinator PUB mechanism optimization (#2442) 
Signed-off-by: NumberWan <wantszkin2003@gmail.com>
 2026-04-24 15:09:23 +08:00
Baoyuan Qi 5632b78b3d [Feature] XiaomiMiMo/MiMo-V2.5-ASR support (#3089) 
Signed-off-by: Baoyuan Qi <qibaoyuan@xiaomi.com>
 2026-04-24 13:55:58 +08:00
LancerandHongsheng Liu 345504b68f [Feat] support quantization for GLM-IMAGE (#2292) 
Signed-off-by: Lancer <maruixiang6688@gmail.com>
Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
 2026-04-24 13:49:49 +08:00
Lancer d20a6b5b39 [chore] refine the examples (#3095) 
Signed-off-by: Lancer <maruixiang6688@gmail.com>
 2026-04-24 13:40:38 +08:00
Alex Brooks 9ebfbf9fde [CI] Re-enable Prefix Cache Test (#2869) 
Signed-off-by: Alex Brooks <albrooks@redhat.com>
 2026-04-24 12:58:25 +08:00
wuhang 2e067bbc24 [Docs] CLI Docs updates (#2978) 
Signed-off-by: wuhang <wuhang6@huawei.com>
 2026-04-24 11:25:22 +08:00
lyj-jjj 19980bb37c [Perf][Wan2.2] Add fused RMSNorm replace WanRMS_norm on npu (#3067) 
Signed-off-by: lyj-jjj <liuyingjun5@huawei.com>
 2026-04-23 18:25:10 +00:00
yuefeng Wu c8efdbd671 [BugFix]: Fix Qwen3-TTS code2wav fails when enforce_eager: false (#2868) 
Signed-off-by: wuyuefeng <565948592@qq.com>
 2026-04-23 09:54:56 -04:00
Hui.andWeiQing Chen 95a07f7732 [PERF]use mindiesd fused rope and rope cache (#2571) 
Signed-off-by: Hui <1779066624@qq.com>
Signed-off-by: Hui. <62495465+Hu1Lcode@users.noreply.github.com>
Co-authored-by: WeiQing Chen <40507679+david6666666@users.noreply.github.com>
 2026-04-23 17:47:43 +08:00
zhumingjue138 47edee1a79 [Test] add stability test case for wan2.2, qwen-tts, qwen3-omni and qwen-image model and modified conftest.py in test/dfx/ (#2817) 
Signed-off-by: zhumingjue <zhumingjue@huawei.com>
Signed-off-by: zhumingjue138 <zhumingjue@huawei.com>
 2026-04-23 17:14:48 +08:00
Ricardo Noriega 05c6c9a0b6 feat: add LTX-2.3 video generation model support (#2893) 
Signed-off-by: Ricardo Noriega De Soto <rnoriega@redhat.com>
 2026-04-23 16:51:51 +08:00
HacoandSYLAR d14ea37842 [BugFix] Preserve media access args for stage configs (#2956) 
Signed-off-by: xiaohajiayou <923390377@qq.com>
Co-authored-by: SYLAR <125541396+lishunyang12@users.noreply.github.com>
 2026-04-23 14:21:37 +08:00
NumberWan 1a7fb55b5f [BugFix] Bagel img2img e2e: drop extra_body height/width (#3054) 
Signed-off-by: NumberWan <wantszkin2003@gmail.com>
 2026-04-23 14:17:25 +08:00
Yuanheng ZhaoLHXuuuHongsheng Liu
 		3a32cd6780 [Model] Ming-flash-omni-2.0 Omni-Speech and TTS (#2890) 
Signed-off-by: Yuanheng Zhao <jonathan.zhaoyh@gmail.com>
Signed-off-by: yuanheng <jonathan.zhaoyh@gmail.com>
Signed-off-by: LHXuuu <xulianhao.xlh@antgroup.com>
Co-authored-by: LHXuuu <xulianhao.xlh@antgroup.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
 2026-04-23 12:04:52 +08:00
WeiQing Chen f327cd1c89 [Bugfix] Fix default sampling params for /v1/videos on main (#3049) 
Signed-off-by: david6666666 <530634352@qq.com>
 2026-04-23 11:31:13 +08:00
WeiQing Chen d4cbdff9cd [Doc] Add diffusion attention backend docs (#3011) 
Signed-off-by: david6666666 <530634352@qq.com>
 2026-04-23 11:30:47 +08:00
R0CKSTAR 8384dead56 [MUSA][Feat] Upgrade MATE to match CUDA/ROCm behavior on FA (#2766) 
Signed-off-by: Xiaodong Ye <yeahdongcn@gmail.com>
 2026-04-23 03:07:02 +00:00
amy-why-3459 3aa84c4742 [Tests]Add accuracy benchmark L4 test cases (#2843) 
Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-04-23 09:40:43 +08:00
wuhang fd9a0ee27f [Feature] Failure message shows more details (#2961) 
Signed-off-by: wuhang <wuhang6@huawei.com>
 2026-04-23 04:32:43 +08:00
Nick CaoandClaude e3b0afbb97 [Refactor] Let diffusion pipelines declare offloadable modules via SupportsModuleOffload (#2427) 
Signed-off-by: Nick Cao <ncao@redhat.com>
Co-authored-by: Claude <noreply@anthropic.com>
 2026-04-23 04:23:00 +08:00
NumberWan ead87aa521 [BugFix] Surface diffusion metrics in chat completions; sanitize Bagel img2img mm kwargs (#2932) 
Signed-off-by: NumberWan <wantszkin2003@gmail.com>
 2026-04-22 16:57:01 +00:00
Zeyu Huang | 黃澤宇 d8cc7a0409 [feat]: General diffusers adapter backend to run diffusion models (#2724) 
Signed-off-by: Huang, Zeyu <11222265+fhfuih@users.noreply.github.com>
 2026-04-23 00:07:18 +08:00
Yueqian Lin 0a43fb94ac style(moss_tts_nano): ruff-format on_requests_finished log call 
Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
 2026-04-22 10:12:07 -04:00
Yueqian Lin 123dfc0772 fix(moss_tts_nano): release stream generators on abnormal termination 
forward() only popped _stream_gens on normal completion, so cancelled,
timed-out, or preempted requests leaked their generator and skipped the
finally block that unlinks the temp WAV files. Implement
on_requests_finished to close each finished generator, which raises
GeneratorExit inside it and triggers the existing cleanup.

Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
 2026-04-22 10:09:58 -04:00
Ricardo Noriega a664e40716 Fix links in glm_image.md (#3028) 
Signed-off-by: Ricardo Noriega <rnoriega@redhat.com>
 2026-04-22 21:08:19 +08:00
Liang LvandSYLAR ee15f39ea7 [AutoRound] Support Qwen Omni W4A16 quantization model (#2670) 
Signed-off-by: lvliang-intel <liang1.lv@intel.com>
Co-authored-by: SYLAR <125541396+lishunyang12@users.noreply.github.com>
 2026-04-22 19:54:45 +08:00
amy-why-3459 9337bec9ad [Tests] Modify test cases (#2991) 
Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-04-22 19:08:10 +08:00
GXINandgcanlin 9d1392dcfc [NPU] Support code predictor NPU graph (#2695) 
Signed-off-by: XIN GAO <1037396230@qq.com>
Signed-off-by: gcanlin <canlinguosdu@gmail.com>
Co-authored-by: gcanlin <canlinguosdu@gmail.com>
 2026-04-22 10:15:08 +00:00
wangyu e18cb89d5a [CI] Update test markers and configurations to use 'full_model' for L4 nightly tests (#2641) 
Signed-off-by: wangyu <410167048@qq.com>
 2026-04-22 15:14:20 +08:00
Jared WenandSYLAR 554233208f [ConfigRefactor] GLM-Image (#2977) 
Signed-off-by: JaredforReal <w13431838023@gmail.com>
Signed-off-by: Jared Wen <w13431838023@gmail.com>
Co-authored-by: SYLAR <125541396+lishunyang12@users.noreply.github.com>
 2026-04-22 15:09:28 +08:00
汪志鹏 e1cdd0c23b [BugFix][CI]: Fix test_omni_sleep_mode ci bug (#3010) 
Signed-off-by: princepride <wangzhipeng628@gmail.com>
 2026-04-22 12:36:20 +08:00
ChenWenjing e82c0a5e1d [ci]add merge ci for streaming input (#2965) 
Signed-off-by: CHEN <116010019@link.cuhk.edu.cn>
 2026-04-22 12:08:56 +08:00
Chen-Yo SunandYueqian Lin 8e1bbc9ab5 [Feature][Voxtral TTS] Support per-model extra sampling params & add cfg_alpha for Voxtral TTS (#2338) 
Signed-off-by: Chen-Yo Sun <chenyo.sun@mistral.ai>
Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
Co-authored-by: Yueqian Lin <linyueqian@outlook.com>
 2026-04-22 03:49:50 +00:00
Greenhand Tsangandgreenhandzpx 2903f528d2 [Bugfix][Model] Fix Qwen3-TTS Code2Wav max_model_len validation (#2508) 
Signed-off-by: greenhandzpx <px.tsang@foxmail.com>
Co-authored-by: greenhandzpx <px.tsang@foxmail.com>
 2026-04-22 03:35:39 +00:00
ChipMates e6cf26efb7 [Qwen3TTS][Bugfix] Guard inner CUDA graph replay during outer capture (#2910) 
Signed-off-by: ChipMates <strasserm@chipmates.ai>
 2026-04-21 23:04:54 -04:00
Divin HonnappaandSYLAR 7a34298743 Remove dead code (#2998) 
Signed-off-by: Divin Honnappa <divin.honnappa@amd.com>
Co-authored-by: SYLAR <125541396+lishunyang12@users.noreply.github.com>
 2026-04-22 08:05:33 +08:00
Yueqian Lin b03a41a7f9 refactor(moss_tts_nano): migrate to Pipeline + Deploy schema 
Follow-up to #2958. MOSS-TTS-Nano was authored before the schema refactor
and still shipped as a legacy stage_configs yaml; this aligns it with the
same layout the 5 migrated TTS models now use.

- vllm_omni/model_executor/models/moss_tts_nano/pipeline.py declares
  MOSS_TTS_NANO_PIPELINE (single LLM_AR stage, owns_tokenizer, audio
  output, stop_token_ids=[2] as a hard EOS backstop).
- vllm_omni/deploy/moss_tts_nano.yaml holds runtime knobs (max_num_seqs,
  gpu_memory_utilization, enforce_eager, default_sampling_params,
  skip_mm_profiling); trust_remote_code stays at deploy top-level.
- vllm_omni/config/pipeline_registry.py registers the entry so the lazy
  registry can resolve it.
- moss_tts_nano/__init__.py exports MossTTSNanoForGeneration (VoxCPM2 pattern).
- Removed vllm_omni/model_executor/stage_configs/moss_tts_nano.yaml.

Examples, shell scripts, READMEs, and buildkite-invoked tests are
updated to use `vllm serve <model> --omni` / `--deploy-config`
(auto-load kicks in; no --stage-configs-path or --trust-remote-code).

Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
 2026-04-21 18:07:50 -04:00
Yueqian Lin 26bbe7b527 fix(moss_tts_nano): address PR review feedback 
- Wire ref_audio through /v1/audio/speech: _build_moss_tts_params is now
  async, resolves ref_audio via MediaConnector and passes the waveform as
  prompt_audio_array so the model can materialise a temp WAV for upstream
  inference_stream (previously voice cloning via the REST endpoint was a
  no-op). Serving now also requires ref_text when ref_audio is provided.
- Fix per-request EOS in batched decode: replace the shared
  _ar_emit_stop_token bool with a _ar_last_chunk_flags list so
  compute_logits emits EOS per row; finished requests no longer wait for
  the slowest peer in a max_num_seqs=4 batch.
- Snapshot and restore CPU+CUDA RNG state around torch.manual_seed to
  limit global-state bleed; add comment noting deterministic output
  under concurrent batching is best-effort (upstream inference_stream
  uses the global RNG).
- Align _DEFAULT_MODE with the offline example and tests ("voice_clone").
- Clean up outdated "buffer first" comment in _create_stream_gen; document
  single-threaded AR-worker assumption for _stream_gens; add one-liner
  explaining the load_weights iterator drain.

Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
 2026-04-21 17:58:50 -04:00