anodyine/vllm-omni
1
0 Fork 0
Code Issues 4 Pull Requests 2 Actions Packages Projects Releases Wiki Activity
1,874 Commits 235 Branches 31 Tags
e592d29f4cdf1a783f0d6b1c634ea6608f89fa92
Commit Graph
1874 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Hongsheng Liu e592d29f4c docs: update README and supported models for v0.22.0 release 
- Add v0.22.0 to Latest News section in README with Cosmos3 recipe link
- Add new models to supported_models.md: MiniCPM-o 4.5, OmniVoice, VoxCPM2,
  MOSS-TTS Nano, MOSS-TTS Delay, MOSS-TTS Realtime, and Lance
- Remove Higgs-Audio v2/v3 (not in 0.22 release)

Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>
 2026-06-08 16:43:49 +08:00
Yueqian Lin 82cce633a3 [Perf][Higgs-Audio-V3] Turn on Stage-0 prefix caching by default (#4199) 
Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
 2026-06-07 03:09:47 +00:00
Yueqian Lin fce0f51089 [Perf][Higgs-Audio-V3] LRU cache for voice-clone ref-audio encode (#4200) 
Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
 2026-06-07 03:09:16 +00:00
Yuekai ZhangandYueqian Lin 82e74ce972 [TTS][New Model] support bosonai/higgs-audio-v3-tts-4b (#4169) 
Signed-off-by: Yuekai Zhang <zhangyuekai@foxmail.com>
Signed-off-by: Yueqian Lin <linyueqian@outlook.com>
Co-authored-by: Yueqian Lin <linyueqian@outlook.com>
 2026-06-06 15:59:12 +00:00
Gao Han 963ba1ab7a [CI] Revert "[Feature] gen-only FP8 Quantization Support for SenseNova-U1" (#4196) v0.22.0 2026-06-06 08:39:10 +08:00
ChenWenjing e25239c2ec [bugfix] fix realtime ci timeout error (#4187) 
Signed-off-by: CHEN <116010019@link.cuhk.edu.cn>
 2026-06-06 00:05:07 +08:00
Zeyu Huang | 黃澤宇andGao Han 2414b6f94d [CI][bugfix]: Improve Qwen Image accuracy test with diffusers' RMSNorm (#4143) 
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>
Signed-off-by: Gao Han <hgaoaf@connect.ust.hk>
Co-authored-by: Gao Han <hgaoaf@connect.ust.hk>
 2026-06-05 07:07:26 -07:00
Chenguang Zheng b3dd0a4baf [Fix] Update Qwen3 Omni multi-replica perf baselines (#4175) 
Signed-off-by: Chenguang ZHENG <645327136@qq.com>
 2026-06-05 09:22:53 +00:00
Jianbao Huang 6cefe91370 [Perf] Enable fused RMSNorm for HunyuanImage3 (#3959) 2026-06-05 07:41:56 +00:00
Yaegaki1Erikaandreandeng 84569280b1 [Bugfix]Fix HunyuanImage3 conditional image and prompt kwargs alignment (#4145) 
Signed-off-by: reandeng <reandeng@tencent.com>
Signed-off-by: Yaegaki1Erika <70182590+Yaegaki1Erika@users.noreply.github.com>
Co-authored-by: reandeng <reandeng@tencent.com>
 2026-06-05 12:47:25 +08:00
amy-why-3459 1e43656ab1 [Bugfix] Update the value of --max-seed-tts-mean-wer in the accuracy test. (#4160) 
Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-06-05 11:39:07 +08:00
Shylin 3875b81c18 [skip ci]cleanup(assets): remove dead vllm_omni/assets/video.py (#4120) 
Signed-off-by: Shylin26 <parishachauhan26@gmail.com>
 2026-06-05 11:21:38 +08:00
Guanzhe (Leo) Huang b63b36ded3 [Feature] gen-only FP8 Quantization Support for SenseNova-U1 (#3943) 
Signed-off-by: leohuang257 <masugchds@gmail.com>
 2026-06-05 02:51:11 +00:00
Yuanheng Zhaoandakshatvishu 95d56cf062 [Bugfix] Fix Compatibility of Ming-flash-omni-2.0 on transformers 5.X and vllm 0.22 (#4080) 
Signed-off-by: Yuanheng Zhao <jonathan.zhaoyh@gmail.com>
Signed-off-by: akshatvishu <akshatnayak197@gmail.com>
Co-authored-by: akshatvishu <akshatnayak197@gmail.com>
 2026-06-05 09:27:42 +08:00
WeiQing Chenandlishunyang12 6e8ef213d8 [BugFix] Support ModelOpt FP8 under batched diffusion serving (#4155) 
Signed-off-by: lishunyang12 <lishunyang12@163.com>
Co-authored-by: lishunyang12 <lishunyang12@163.com>
 2026-06-05 06:25:01 +08:00
Chendi.XueandMengkejiergeli Ba 31f882935d [XPU][S2V] Optimize Wan2.2 S2V: RoPE refactor + cache_dit enabling (#4062) 
Signed-off-by: Chendi Xue <chendi.xue@intel.com>
Co-authored-by: Mengkejiergeli Ba <mengkejiergeli.ba@intel.com>
 2026-06-05 06:20:16 +08:00
Zheng Wengang fccd95b12b [BugFix] Fix incorrect GPU device mapping in multi-replica stages (#4132) 
Signed-off-by: ZhengWG <zwg0606@gmail.com>
 2026-06-04 15:32:58 +00:00
dengyunyang 469c1dc87e [HunyuanImage3.0][Performance][Optimazation]Adjust perf config (#4149) 
Signed-off-by: dengyunyang <584797741@qq.com>
 2026-06-04 22:54:46 +08:00
amy-why-3459 519cf3ecf1 [Refactor] [Qwen3-Omni]Modify the thinker's sampling parameters to align with transformers. (#4137) 
Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-06-04 22:49:04 +08:00
ZacheryAU bfc7cedcf1 Benchmark data statistics for each stage of omni models (#3628) 
Signed-off-by: ZacheryAU <70869627+ZacheryAU@users.noreply.github.com>
 2026-06-04 21:43:35 +08:00
Alex Brooks e10aca30ec [CI] Update Bagel Pixels (#4081) 
Signed-off-by: Alex Brooks <albrooks@redhat.com>
 2026-06-04 10:37:34 +00:00
Alicia c6e8d78eda Add dependency FlagEmbedding (#3980) 
Signed-off-by: Alicia <115451386+congw729@users.noreply.github.com>
 2026-06-04 18:00:12 +08:00
BruceLoveDecimal e5030c7524 [BugFix] Cast bf16 video frames to float32 before .numpy() in /v1/videos (#4114) 
Signed-off-by: BruceLoveDecimal <39156883+BruceLoveDecimal@users.noreply.github.com>
 2026-06-04 17:49:07 +08:00
Yuekai Zhang a1cc583665 fix: cosyvoice3 batch>1 inference (#3910) 
Signed-off-by: Yuekai Zhang <zhangyuekai@foxmail.com>
 2026-06-04 08:58:53 +00:00
dengyunyang a57b8f858b [HunyuanImage3][CI] fix ci (#4134) 
Signed-off-by: dengyunyang <584797741@qq.com>
 2026-06-04 16:15:49 +08:00
herotai214 9a9c8f7503 [Test] Add L4 diffusion feature test for GLM-Image (#3451) 
Signed-off-by: herotai214 <herotai214@gmail.com>
 2026-06-04 00:11:17 -07:00
Zhou Taichang af02713952 [Bugfix] harden diffusion model prefetch against transformers v5 shard-resolution race (#4076) 2026-06-04 12:14:25 +08:00
tanhaoan333 d8dc1defa7 [NPU][BugFix] Upgrade parts of ModelRunner to v0.22.0 (#4130) 
Signed-off-by: tanhaoan333 <tanhaoan@huawei.com>
 2026-06-04 03:59:26 +00:00
tanhaoan333 800c2b8251 [bugfix]qwen3tts code2wav (#4123) 
Signed-off-by: tanhaoan333 <tanhaoan@huawei.com>
 2026-06-04 02:20:00 +00:00
Canlin Guo 221584a489 [CI] Skip online moss test temp (#4122) 2026-06-04 01:56:11 +00:00
amy-why-3459 4e1c13236e [Perf][Qwen3-Omni]Optimize TTFP using initial_codec_chunk_frames (#4054) 
Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-06-04 09:08:13 +08:00
bastefaniakMaciej Balalishunyang12SYLAR
 		706bad253d Add Cosmos3 action modality (#4102) 
Signed-off-by: Maciej Bala <mbala@nvidia.com>
Signed-off-by: Bartosz Stefaniak <bstefaniak@nvidia.com>
Signed-off-by: lishunyang12 <lishunyang12@163.com>
Co-authored-by: Maciej Bala <mbala@nvidia.com>
Co-authored-by: lishunyang12 <lishunyang12@163.com>
Co-authored-by: SYLAR <125541396+lishunyang12@users.noreply.github.com>
 2026-06-03 19:30:39 +00:00
amy-why-3459 8949fdc440 [BugFix] Fix the issue of vllm failing to start. (#4105) 
Signed-off-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-06-03 15:44:46 +00:00
Canlin Guo 3818ba4c65 [CI] Remove omni mark for MOSS-TTS and temporarily skipped (#4097) 
Signed-off-by: gcanlin <canlinguosdu@gmail.com>
 2026-06-03 23:12:11 +08:00
Sy03Yueqian LinHongsheng LiuGao Han
 		c33ca60208 [Perf][VoxCPM2] Optimize VoxCPM2 high-concurrency decode throughput. (#3882) 
Signed-off-by: Sy03 <1370724210@qq.com>
Co-authored-by: Yueqian Lin <70319226+linyueqian@users.noreply.github.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
Co-authored-by: Gao Han <hgaoaf@connect.ust.hk>
 2026-06-03 23:08:32 +08:00
tanhaoan333 0fc4032a71 Update qwen3_tts_code2wav.py (#4075) 
Signed-off-by: tanhaoan333 <tanhaoan@huawei.com>
 2026-06-03 13:34:47 +00:00
ChenWenjing 2e108a4109 [ci] add merge/ready ci for audio realtime api (#4069) 
Signed-off-by: CHEN <116010019@link.cuhk.edu.cn>
 2026-06-03 18:57:16 +08:00
Canlin Guo ce69eb35e9 [Refactor] Refactor HunyuanImage3 SigLIP2 ViT to vLLM layers (#3297) 
Signed-off-by: gcanlin <canlinguosdu@gmail.com>
 2026-06-03 17:51:06 +08:00
Ricardo NoriegaClaude Opus 4.6Alex Brooks
 		07737e6c61 [Test] Add prefix caching + audio output regression test (#3510) (#3604) 
Signed-off-by: Ricardo Noriega De Soto <rnoriega@redhat.com>
Signed-off-by: Alex Brooks <albrooks@redhat.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: Alex Brooks <albrooks@redhat.com>
 2026-06-03 09:17:53 +00:00
zzhAliciaGao HanHongsheng Liu
 		d2e65a92f1 [CI][Accuracy] Add HunyuanImage3 pixel accuracy test and nightly CI (#3790) 
Signed-off-by: zzh <943967662@qq.com>
Co-authored-by: Alicia <115451386+congw729@users.noreply.github.com>
Co-authored-by: Gao Han <hgaoaf@connect.ust.hk>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
 2026-06-03 16:59:21 +08:00
ChenWenjing 3842970a82 [bugfix] fix streaming input issue after rebase 0.22.0 (#4085) 
Signed-off-by: CHEN <116010019@link.cuhk.edu.cn>
 2026-06-03 16:07:41 +08:00
amy-why-3459 9bb4974198 [BugFix] Fix the issue of dataset names not being resolved (#4094) 2026-06-03 06:17:04 +00:00
ChenWenjing 0e2d55f46d [ci] add Voxcpm2 accuracy tests (#4055) 
Signed-off-by: CHEN <116010019@link.cuhk.edu.cn>
 2026-06-03 14:11:04 +08:00
MaciejBalaNVBartosz Stefaniaklishunyang12
 		40b2959183 Add Cosmos3 sound generation (#4073) 
Signed-off-by: Maciej Bala <mbala@nvidia.com>
Signed-off-by: lishunyang12 <lishunyang12@163.com>
Signed-off-by: Bartosz Stefaniak <bstefaniak@nvidia.com>
Co-authored-by: Bartosz Stefaniak <bstefaniak@nvidia.com>
Co-authored-by: lishunyang12 <lishunyang12@163.com>
 2026-06-03 00:51:44 +00:00
Honghan ZhuandClaude Opus 4.8 3329f1b674 [Refactor] Migrate dynin_omni to pipeline registry, drop legacy stage… (#4078) 
Signed-off-by: dph97 <nenbaying@163.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
 2026-06-02 21:28:12 +00:00
Nick CaoandClaude 23484d8a11 [Bugfix] Pass media paths to use_mixed_modalities in example script (#3355) 
Signed-off-by: Nick Cao <ncao@redhat.com>
Co-authored-by: Claude <noreply@anthropic.com>
 2026-06-02 21:04:16 +00:00
Yangshen DengMengSuli Wangamy-why-3459
 		bdfc77147d [Diffusion] DreamZero world model integration with CFG parallel + OpenPI serving (#2162) 
Signed-off-by: Meng <meng_chen99@163.com>
Signed-off-by: Yangshen Deng <yangshen.d@outlook.com>
Signed-off-by: Suli Wang <50828270+wsl2000@users.noreply.github.com>
Co-authored-by: Meng <meng_chen99@163.com>
Co-authored-by: Suli Wang <50828270+wsl2000@users.noreply.github.com>
Co-authored-by: amy-why-3459 <wuhaiyan17@huawei.com>
 2026-06-02 19:50:20 +00:00
liqianGao HanWeiQing Chen
 		e35cd213d7 [NPU] [Perf] Adjust flash_attn mask shape for hunyuanvideo1.5 on npu (#3178) 
Signed-off-by: liqian <65649795+vasede@users.noreply.github.com>
Co-authored-by: Gao Han <hgaoaf@connect.ust.hk>
Co-authored-by: WeiQing Chen <40507679+david6666666@users.noreply.github.com>
 2026-06-02 23:23:18 +08:00
NumberWan 35ee3c7543 [BugFix]Qwen-Image performance regression by using torch RMSNorm(RMSNorm backend) (#4074) 
Signed-off-by: NumberWan <wantszkin2003@gmail.com>
 2026-06-02 23:19:17 +08:00
rein yangandHongsheng Liu 3a7c7f1422 [Bugfix] fix qwen3-omni performance regression (#3575) 
Signed-off-by: rein yang <ruiruyang2@gmail.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
 2026-06-02 23:02:07 +08:00