Commits · ab1a4fa5cbb0cf5d2ea9a2f3b36f16fd129cf945 · change / sglang

"git@developer.sourcefind.cn:OpenDAS/mmdetection3d.git" did not exist on "b77a77d4b9d360a7f5e73841a91a6909a8474f40"

14 Jun, 2025 7 commits
- [fix] fix cutlass_mla_backend with cuda_graph and add sm_scale for sgl-kernel cutlass_mla (#7184) · ab1a4fa5
  JieXin Liang authored Jun 15, 2025
  
  ab1a4fa5
- [fix] fix dsv3 weight loader tqdm and simplify shared experts fusion (#7181) · ed54bf9d
  JieXin Liang authored Jun 15, 2025
  
  ed54bf9d
- Fix shared experts fusion + weight requant (#7177) · b57d87c2
  fzyzcjy authored Jun 14, 2025
  
  b57d87c2
- Increase timeout in test/srt/test_disaggregation.py (#7175) · f47a1b1d
  Lianmin Zheng authored Jun 13, 2025
  
  f47a1b1d
- Support new DeepGEMM (#7172) · 93cec433
  fzyzcjy authored Jun 14, 2025
  
  93cec433
- Refactor DeepGEMM integration (#7150) · b4c41f72
  fzyzcjy authored Jun 14, 2025
  
  b4c41f72
- Support new DeepGEMM input format in silu_and_mul_masked_post_quant_fwd (#7153) · 8b8f2e74
  fzyzcjy authored Jun 14, 2025
  
  8b8f2e74
13 Jun, 2025 10 commits
- Minor style change of triton backend (#7165) · be2d985d
  Lianmin Zheng authored Jun 13, 2025
  
  be2d985d
- Re-quantize DeepSeek model weights to support DeepGEMM new input format (#7156) · 5b1afa78
  fzyzcjy authored Jun 14, 2025
  
  5b1afa78
- Remove 200us slow concat kernel (part 2: srt) (#7020) · c49c1d92
  fzyzcjy authored Jun 14, 2025
  
  c49c1d92
- Tiny add sanity checks for DeepGEMM inputs (#7157) · 0f1dfa1e
  fzyzcjy authored Jun 14, 2025
  
  0f1dfa1e
- Minor speed up block_quant_dequant (#6814) · e3ec6bf4
  fzyzcjy authored Jun 14, 2025
  
  e3ec6bf4
- Fix DeepEP error in some environments (#7154) · b04df75a
  fzyzcjy authored Jun 14, 2025
  
  b04df75a
- Support new DeepGEMM format in per token group quant (part 2: srt) (#7155) · bec3e484
  fzyzcjy authored Jun 14, 2025
  
  bec3e484
- filter by num_hidden_layers (#7056) · 2f4ec752
  pansicheng authored Jun 13, 2025
  
  2f4ec752
- Minor speedup topk postprocessing (#7058) · da47621c
  fzyzcjy authored Jun 13, 2025
  
  da47621c
- Remove unnecessary metadata_expand.max_seq_len_k operations in fa3 to… (#7140) · 22a6b9fc
  Binyao Jiang authored Jun 12, 2025
  
  22a6b9fc
12 Jun, 2025 4 commits
- Minor add utility to read expert distribution recorder output (#7134) · b02df20a
  fzyzcjy authored Jun 13, 2025
  
  b02df20a
- [Fix] Reduce busy polling when scheduler is idle (#6026) · bd7cfbd2
  Povilas Kanapickas authored Jun 13, 2025
  
  bd7cfbd2
- Improve log status (#7115) · 930746d9
  Liangsheng Yin authored Jun 12, 2025
  
  930746d9
- Fix positional argument (#7093) · ef326774
  Quanfeng Li authored Jun 12, 2025
  
  ef326774
11 Jun, 2025 11 commits
- [Perf] Refactor LoRAManager to eliminate stream syncs and redundant computations (#6994) · 021f76e4
  Lifu Huang authored Jun 11, 2025
  
  021f76e4
- [feat]: Emit fixed-size KV blocks events (#6824) · 777688b8
  Faradawn Yang authored Jun 11, 2025
  
  777688b8
- [FIX]remove redundant code in logits_processor.py (#7079) · 0ca594ed
  Neo authored Jun 12, 2025
  
  0ca594ed
- Support VILA models (#6106) · 31d6dee5
  Zijian authored Jun 12, 2025
  
  31d6dee5
- Fix misusing the "_is_cuda". (#7091) · 02543b54
  sogalin authored Jun 12, 2025
  
  02543b54
- Fix circular import in test_prefix_chunk_info.py (#7097) · 25a6a9aa
  Baizhou Zhang authored Jun 11, 2025
  
  25a6a9aa
- vlm: adapt internvl to VisionAttention (#6870) · 83d87685
  Mick authored Jun 11, 2025
  
  83d87685
- Fix GGuf and add back test_gguf.py (#7067) · 2a5f0100
  Baizhou Zhang authored Jun 10, 2025
  
  2a5f0100
- Clean up docs for server args and sampling parameters (generated by grok) (#7076) · dbdf76ca
  Lianmin Zheng authored Jun 10, 2025
  
  dbdf76ca
- Simplify the heuristics for setting --mem-fraction-static (#7054) · 6b12d6a8
  Lianmin Zheng authored Jun 10, 2025
  
  6b12d6a8
- Frontend language separate reasoning support (#6031) · 14c18d25
  Yudi Xue authored Jun 10, 2025
  
  14c18d25
10 Jun, 2025 8 commits
- [Feature] Add Logit Bias (#6579) · ca929118
  Brayden Zhong authored Jun 10, 2025
```
Co-authored-by: Cinjon Resnick <cinjon.resnick@gmail.com>
```
  ca929118
- Open AI API hidden states (#6716) · b56de8f9
  kyle-pena-kuzco authored Jun 10, 2025
  
  b56de8f9
- Fix missing tool call id if tool call index >0 in streaming tool call output. (#7049) · a0e4d4eb
  Xu Wenqing authored Jun 11, 2025
```
Signed-off-by: 许文卿 <xwq391974@alibaba-inc.com>
```
  a0e4d4eb
- [WA] fix output data is nan in CI test "test_moe_eval_accuracy_large.py" (#7021) · 8ea7df61
  kk authored Jun 11, 2025
```
Co-authored-by: wunhuang <wunhuang@amd.com>
Co-authored-by: HAI <hixiao@gmail.com>
```
  8ea7df61
- Minor style fix in cuda_graph_runner.py (#7053) · 4a102a2b
  Lianmin Zheng authored Jun 10, 2025
  
  4a102a2b
- Clean up server_args.py (#7037) · 6406408a
  Lianmin Zheng authored Jun 10, 2025
  
  6406408a
- Fix eagle on AMD (#7051) · 019851d0
  Lianmin Zheng authored Jun 10, 2025
  
  019851d0
- Minor cleanup of fa3 backend (#6999) · 2dae104d
  Lianmin Zheng authored Jun 10, 2025
  
  2dae104d