anodyine/vllm-omni
1
0 Fork 0
Code Issues 4 Pull Requests 2 Actions Packages Projects Releases Wiki Activity
2,942 Commits 235 Branches 31 Tags
c4e61133fbb5587e0cf8972ce32ffac4e37fa739
Commit Graph
32 Commits
Author SHA1 Message Date
tzhouamandClaude Fable 5.1 9502d3cef7 [Edge][CPU] Compiled post-step also covers the Identity talker->predictor projection (0.6B) 
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
 2026-09-10 03:48:25 +08:00
tzhouamandClaude Fable 5.1 3085f50160 [Edge][CPU] Compile the per-sub-step lm head + top-k Gumbel sampling + embedding + projection into one graph 
Noise is drawn eagerly with the request generators (same RNG consumption as the
eager loop) and passed in, so the compiled post-step is generator-free and
shape-static; used on the fp32/torchao CPU path outside stored (top-p) mode.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
 2026-09-10 03:12:40 +08:00
tzhouamandClaude Fable 5.1 82955b3a22 [Edge][CPU] Default the code predictor's CPU fast path to fp32 + compiled static step; int8 opt-in 
In-situ parity against the bf16 model (VLLM_OMNI_PREDICTOR_PARITY_CHECK=1):
fp32 control keeps 93.8 % top-1 / 95 % top-5 / KL 0.0015; torch.ao dynamic
int8 falls to 22 % top-1 (max logit diff 7), so per-tensor activation
quantization is not acceptable as a default. Adds the ao-nohead variant
(int8 transformer body, fp32 lm heads) for measurement.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
 2026-09-10 02:54:53 +08:00
tzhouamandClaude Fable 5.1 94485152d2 [Edge][CPU] Predictor fast-path modes: torch.ao int8 (default), torchao weight-only + compiled step, fp32 control; parity metrics 
torchao's dynamic-activation int8 is numerically wrong under torch.compile on
CPU (cosine 0.02 vs fp32; caused runaway generations), so it is not offered;
weight-only + compile is faithful but slower (78 ms vs 36 ms per frame for the
15 sub-steps), so torch.ao dynamic int8 stays the default. Parity mode now also
records top-5 overlap and KL vs the bf16 model; compiled-step equivalence test.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
 2026-09-10 02:44:29 +08:00
tzhouamandClaude Fable 5.1 6ce8de1b28 [Edge][CPU] KV-cached code predictor, int8-dynamic CPU fast path, model-side CPU sampler 
Measured on a Xeon 8480C (12+4 threads, 0.6B): the 15 residual-codebook sub-steps
re-ran the 5-layer predictor over the whole prefix each time (97 ms per frame of
the 134 ms talker step) and vLLM's stock sampler spent 9 ms per step in its
torch.compile'd random-sample kernel on CPU.

- code predictor: forward_cached (KV cache per layer, 2-token prefill then one
  token per sub-step) and a shape-static forward_static_step (tensor position +
  boolean mask) that torch.compile handles as one graph; incremental decoding
  is the default on CPU (VLLM_OMNI_PREDICTOR_INCREMENTAL), optional compile
  (VLLM_OMNI_PREDICTOR_COMPILE=step), optional thread cap.
- CPU fast path: dynamic int8 Linear layers (fp32 activations) for the predictor
  and its lm heads (VLLM_OMNI_PREDICTOR_INT8, default on; micro-benchmark 70 ->
  36 ms per frame for the 15 sub-steps); VLLM_OMNI_PREDICTOR_PARITY_CHECK=1
  records top-1 agreement / max logit diff against the bf16 model in situ.
- Qwen3-TTS talker: prefer_model_sampler on CPU with a sample() that mirrors
  vLLM's Sampler (logits processors + penalties, temperature, top-k/top-p,
  exponential-noise sampling with per-request generators) without the compiled
  kernel (VLLM_OMNI_CPU_FAST_SAMPLER, default on).
- step_stats: sampler phase timers and sample-metadata timers.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
 2026-09-10 01:56:00 +08:00
tzhouamandClaude Fable 5.1 abf81f4114 [Edge][WP8] Model-level step timers: talker forward, sampling, talker-MTP, code-predictor sub-steps 
Counters model.forward_ms / model.sample_ms / model.talker_mtp_ms and
predictor.substep_ms / predictor.loop_ms (VLLM_OMNI_STEP_STATS_DIR); with
VLLM_OMNI_STEP_STATS_SYNC=1 the accelerator is synchronized around the model
timers so async kernels are attributed (attribution runs only).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01L1ax5wYLA6XbqEL1i8L9gZ
 2026-09-09 21:32:15 +08:00
tzhouamandClaude Fable 5.1 073d31b9cd [Edge][WP2] Add initialization timeline and cold-start profiler 
* vllm_omni/engine/init_timeline.py: InitTimeline phase collector; engine
  phases (stages_init, g3_device_lock, g2_spawn/ready) and worker phases
  (proc_entry, engine_core_init, init_device, load_model/load_weights,
  talker_mtp_init, attn_prewarm, cudagraph_capture, talker_mtp_capture,
  code2wav_cudagraph, predictor_setup_compile/warmup/cudagraph) recorded via
  worker_phase/timed_phase; workers append JSONL under
  VLLM_OMNI_INIT_TIMELINE_DIR (inherited at spawn) and the engine merges them.
  Enabled by VLLM_OMNI_INIT_TIMELINE=<json> or log_stats; table logged at INFO,
  partial table on orchestrator startup timeout, suggested timeouts logged.
* omni_base.py: named deploy profiles default to stage_init_timeout=900 /
  init_timeout=1500 unless given explicitly.
* gpu_model_runner._init_talker_mtp tolerates max_cudagraph_capture_size=None;
  gpu_ar_model_runner.shutdown skips CUDA graph teardown on CPU.
* benchmarks/tts/init_profile.py: runs stream_latency_bench --init-only per
  configuration (default, cold_cache, eager_stage1, eager_both,
  parallel_stage_init, edge, ...) and tabulates phase durations.
* stream_latency_bench: --parallel-stage-init / --deploy-profile flags; the
  resource poller no longer shadows Thread._stop and skips GPU sampling on CPU.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
 2026-09-09 11:01:39 +08:00
Joshna-MedisettyandCursor 24c49cf392 [Bugfix][XPU] Add MoTRMSNorm.forward_xpu delegating to forward_native (#5568) 
Signed-off-by: Joshna Medisetty <joshna.medisetty@intel.com>
Signed-off-by: Joshna-Medisetty <joshna.medisetty@intel.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
 2026-09-02 23:06:17 -05:00
林鑫sheenlinyurainGao Han
 		37c11e6e4f [Perf][Qwen3-TTS] Fuse QKV and gate_up projections in code predictor (#5791) 
Signed-off-by: sheenlin <sheenlin@tencent.com>
Signed-off-by: Gao Han <hgaoaf@connect.ust.hk>
Co-authored-by: sheenlin <sheenlin@tencent.com>
Co-authored-by: yurain <yurain26@users.noreply.github.com>
Co-authored-by: Gao Han <hgaoaf@connect.ust.hk>
 2026-08-20 19:22:40 +00:00
dengyunyang 8d34d3b5de [Perf][Hunyuan] Optimize for vae and patch_embed (#6306) 
Signed-off-by: dengyunyang <584797741@qq.com>
 2026-08-20 16:29:21 +00:00
Yuanheng Zhao a630b0c2ec [Ming] Cleanup Ming-family shared modules (#6119) 
Signed-off-by: Yuanheng <jonathan.zhaoyh@gmail.com>
Signed-off-by: Yuanheng Zhao <jonathan.zhaoyh@gmail.com>
 2026-08-19 14:13:40 +08:00
Yueqian Lin b283215c28 [Model] Add MiniMax Music 3 text-to-music (#6186) 
Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
 2026-08-18 04:33:13 +00:00
Sy03andYueqian Lin 58007c0084 [New Model] Support IndexTTS 2.5 (#5957) 
Signed-off-by: Sysy <1370724210@qq.com>
Co-authored-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
 2026-08-11 23:00:35 +00:00
Gao Han 78c144f3a8 [CI failed] Revert "[Perf][Qwen3-TTS] Fuse QKV and gate_up projections in code predictor" (#5777) 2026-08-04 15:21:38 +00:00
林鑫sheenlinyurain
 		6fdc4ca62d [Perf][Qwen3-TTS] Fuse QKV and gate_up projections in code predictor (#4958) 
Signed-off-by: sheenlin <sheenlin@tencent.com>
Co-authored-by: sheenlin <sheenlin@tencent.com>
Co-authored-by: yurain <yurain26@users.noreply.github.com>
 2026-08-04 11:59:32 +00:00
林鑫 d33c905d80 [Perf] Cache RoPE cos/sin tables in code predictor (#5503) 
Signed-off-by: I-wave <2083236231@qq.com>
 2026-07-31 22:03:18 +08:00
zyz111222andamy-why-3459 2d58070f60 [Hardware][Ascend][Model] Optimize Qwen3-TTS on 310P (#4841) 
Signed-off-by: zouyizhou <zouyizhou@huawei.com>
Co-authored-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-07-15 08:57:44 +08:00
Zheng Wengang 3f4abedbda [Perf] Ming-TTS Stage-0 flow-head speedup: RoPE/QKV fusion, opt-in torch.compile, PIECEWISE cudagraph (#4942) 2026-07-12 19:50:51 -07:00
林鑫andyurain 020792e6be [Perf][Qwen3-TTS] Sample code predictor outputs via Gumbel-max (#4261) 
Signed-off-by: 林鑫 <77673148+l-wave@users.noreply.github.com>
Co-authored-by: yurain <yurain26@users.noreply.github.com>
 2026-07-10 05:41:54 +00:00
Yueqian Lin d7b365df64 [Bugfix][Qwen3-TTS] Keep seeded residual MTP sampling batched (#4889) 
Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
 2026-07-05 21:47:06 +08:00
LHXuuu 991fb7408b [BugFix][Ming-TTS] Improve Seed-TTS WER accuracy (#4859) 
Signed-off-by: LHXuuu <xulianhao.xlh@antgroup.com>
 2026-07-03 12:15:03 +02:00
LHXuuuYuanheng ZhaoYuanheng ZhaoHongsheng Liu
 		0b04f2a0d6 [Model] Add Ming-omni-tts MoE 16.8B-A3B + CFM CUDAGraph (#4341) 
Signed-off-by: LHXuuu <xulianhao.xlh@antgroup.com>
Signed-off-by: Yuanheng Zhao <jonathan.zhaoyh@gmail.com>
Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>
Co-authored-by: Yuanheng Zhao <jonathan.zhaoyh@gmail.com>
Co-authored-by: Yuanheng Zhao <54058983+yuanheng-zhao@users.noreply.github.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
 2026-06-25 22:11:35 +08:00
akshatvishuandHongsheng Liu 2fa4c7a4a9 [Model] Add Ming-omni-tts dense 0.5B pipeline (#2906) 
Signed-off-by: akshatvishu <akshatnayak197@gmail.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
 2026-06-10 20:11:04 +08:00
BeatSeat dab596e468 [Refactor] Unify Snake/SnakeBeta and alias-free activation into common modules (#3886) 
Signed-off-by: BeatSeat <wendavid552@gmail.com>
 2026-05-27 23:41:35 -05:00
BeatSeatandYueqian Lin 28ce618f0e [Model] Add GLM-TTS text-to-speech model support (#3141) 
Signed-off-by: BeatSeat <wendavid552@gmail.com>
Signed-off-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
Co-authored-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
 2026-05-24 05:42:26 +08:00
Sy03Hongsheng LiuSYLAR
 		c99df1ebd9 [TTS][Perf] Optimize Qwen3-TTS high-concurrency serving (#3662) 
Signed-off-by: Sy03 <1370724210@qq.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
Co-authored-by: SYLAR <125541396+lishunyang12@users.noreply.github.com>
 2026-05-17 16:58:59 -04:00
+20
Zhou TaichangClaude Opus 4.7NumberWandsinghviDnoobYueqian LinHongsheng Liuknlnguyen1802Samitrein yangNick Caozhumingjue138Ricardo Noriegalyj-jjjCursorgcanlinwangyuweizhoublueweizhoubluedengyunyang汪志鹏WeiQing ChenJie LiuYueqian LinNATUREbjf-frzamy-why-3459Y. FisherZeyu Huang | 黃澤宇Copilot Autofix powered by AI
 		eb4e60ee64 [Rebase] Rebase to vllm v0.21.0 (#3530) 
Signed-off-by: tzhouam <tzhouam@connect.ust.hk>
Signed-off-by: Zhou Taichang <tzhouam@connect.ust.hk>
Signed-off-by: NumberWan <wantszkin2003@gmail.com>
Signed-off-by: Divyansh Singhvi <divyanshsinghvi@gmail.com>
Signed-off-by: dsinghvi <divyanshsinghvi@gmail.com>
Signed-off-by: Dnoob <dxpouo@gmail.com>
Signed-off-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
Signed-off-by: knlnguyen1802 <knlnguyen1802@gmail.com>
Signed-off-by: rein yang <ruiruyang2@gmail.com>
Signed-off-by: Nick Cao <ncao@redhat.com>
Signed-off-by: zhumingjue <zhumingjue@huawei.com>
Signed-off-by: Ricardo Noriega De Soto <rnoriega@redhat.com>
Signed-off-by: lyj-jjj <liuyingjun5@huawei.com>
Signed-off-by: gcanlin <canlinguosdu@gmail.com>
Signed-off-by: wangyu <410167048@qq.com>
Signed-off-by: weizhoublue <weizhoublue@github.com>
Signed-off-by: weizhou.lan@daocloud.io <weizhou.lan@daocloud.io>
Signed-off-by: dengyunyang <584797741@qq.com>
Signed-off-by: 汪志鹏 <wangzhipeng628@gmail.com>
Signed-off-by: David Chen <530634352@qq.com>
Signed-off-by: Jie Liu <33612777+keeper-jie@users.noreply.github.com>
Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
Signed-off-by: princepride <wangzhipeng628@gmail.com>
Signed-off-by: natureofnature <wzliu@connect.hku.hk>
Signed-off-by: bjf-frz <frz123db@gmail.com>
Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>
Signed-off-by: KexiongYu <yukexiong1@huawei.com>
Signed-off-by: Huang, Zeyu <11222265+fhfuih@users.noreply.github.com>
Signed-off-by: Zeyu Huang | 黃澤宇 <11222265+fhfuih@users.noreply.github.com>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Co-authored-by: NumberWan <wantszkin2003@gmail.com>
Co-authored-by: dsinghvi <divyanshsinghvi@gmail.com>
Co-authored-by: Dnoob <dxpouo@gmail.com>
Co-authored-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
Co-authored-by: knlnguyen1802 <knlnguyen1802@gmail.com>
Co-authored-by: Samit <285365963@qq.com>
Co-authored-by: rein yang <73573651+R2-Y@users.noreply.github.com>
Co-authored-by: Nick Cao <ncao@redhat.com>
Co-authored-by: zhumingjue138 <zhumingjue@huawei.com>
Co-authored-by: Ricardo Noriega <rnoriega@redhat.com>
Co-authored-by: lyj-jjj <liuyingjun5@huawei.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: gcanlin <canlinguosdu@gmail.com>
Co-authored-by: wangyu <53896905+yenuo26@users.noreply.github.com>
Co-authored-by: weizhoublue <45163302+weizhoublue@users.noreply.github.com>
Co-authored-by: weizhoublue <weizhoublue@github.com>
Co-authored-by: dengyunyang <584797741@qq.com>
Co-authored-by: 汪志鹏 <wangzhipeng628@gmail.com>
Co-authored-by: WeiQing Chen <40507679+david6666666@users.noreply.github.com>
Co-authored-by: Jie Liu <33612777+keeper-jie@users.noreply.github.com>
Co-authored-by: Yueqian Lin <linyueqian@outlook.com>
Co-authored-by: NATURE <wzliu@connect.hku.hk>
Co-authored-by: bjf-frz <frz123db@gmail.com>
Co-authored-by: amy-why-3459 <wuhaiyan17@huawei.com>
Co-authored-by: Y. Fisher <yukexiong1@huawei.com>
Co-authored-by: Zeyu Huang | 黃澤宇 <11222265+fhfuih@users.noreply.github.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
 2026-05-16 18:18:43 +08:00
Juan Pablo ZuluagaandYueqian Lin 19f8f42822 [Feat][Qwen3-Omni] Add CUDA graph support for Code2Wav decoder (#2376) 
Signed-off-by: JuanPZuluaga <juanz9312@gmail.com>
Co-authored-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
 2026-05-07 06:16:03 +08:00
Sy03 2b85474c23 [Bugfix] Propagate seed to Qwen3-TTS Fast AR sampler (#3350) 
Signed-off-by: Sy03 <1370724210@qq.com>
 2026-05-05 14:57:55 -04:00
Nick CaoandClaude b9789091e9 [Bugfix][Qwen3TTS] Use float32 for code predictor on fp16-only GPUs (#3253) 
Signed-off-by: Nick Cao <ncao@redhat.com>
Co-authored-by: Claude <noreply@anthropic.com>
 2026-04-29 23:25:38 +00:00
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
Juan Pablo ZuluagaandHongsheng Liu 82f8c93343 [Feat][Qwen3-Omni] Shared code predictor module for Qwen3-TTS and Qwen3-Omni (#2375) 
Signed-off-by: JuanPZuluaga <juanz9312@gmail.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
 2026-04-15 15:14:57 +08:00