Commits

Branch selector

 develop

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 14, 2026

• [RL][CI] Support Async R3 And Add Accuracy Test (#5937)

  Show description for b3f59fd
  gongshaotian

  authored
  b3f59fd

  Copy full SHA for b3f59fd

• [XPU] fix multi-batch bug in VL model (#6015)

  Show description for 9373f37
  

  ddchenhao66

  and

  plusNew001

  authored
  9373f37

  Copy full SHA for 9373f37

• [Optimize] Qwen2.5-VL vision model with merged linear layers and unif… (#6037)

  Show description for 6f72be7
  xiaoxiaohehe001

  authored
  6f72be7

  Copy full SHA for 6f72be7

• [Feature]Report FD statistical information (#5646)

  Show description for 93b7675
  luukunn

  authored
  93b7675

  Copy full SHA for 93b7675

• [Metax][Fix] fix self.share_inputs['preempted_idx']=[] incorrect use (#6038)

  Show description for 273e79a
  

  StareAtYou

  and

  root

  authored
  273e79a

  Copy full SHA for 273e79a

• [Metax][Doc] update metax gpu 'get_started' doc (#6035)

  Show description for 32fb047
  

  StareAtYou

  and

  root

  authored
  32fb047

  Copy full SHA for 32fb047

• [CI] Adapt vl_model baseline changes due to Paddle update_2 (#6033)

  EmmonsCurse

  authored
  2c17acd

  Copy full SHA for 2c17acd

• [Metax][CI] remove 28B VL model test sampling randomness (#6032)

  Show description for f3587b5
  

  StareAtYou

  and

  root

  authored
  f3587b5

  Copy full SHA for f3587b5

• [XPU][CI] Cache queue port bug fix (#6030)

  Show description for 926a260
  plusNew001

  authored
  926a260

  Copy full SHA for 926a260

• [Optim] Robust sync status when preempted happens (#5796)

  Show description for 74d0f1c
  

  rainyfly

  and

  EmmonsCurse

  authored
  74d0f1c

  Copy full SHA for 74d0f1c

• [Graph Optimization] Add `full_cuda_graph` to control subgraph split (#6027)

  DrRyanHuang

  authored
  0d1a5e7

  Copy full SHA for 0d1a5e7

Commits on Jan 13, 2026

• [BugFix] fix cache transfer manager updating/clearing (#5930)

  Show description for 4566370
  liyonghua0910

  authored
  4566370

  Copy full SHA for 4566370

• [Featue] Enable output caching by default (#5987)

  Show description for 6da06ab
  

  rainyfly

  and

  Jiang-Jia-Jun

  authored
  6da06ab

  Copy full SHA for 6da06ab

• [Metax][CI] update test_ernie_28b_vl.py image result keywords (#6022)

  Show description for 3772810
  

  StareAtYou

  and

  root

  authored
  3772810

  Copy full SHA for 3772810

• [Metax][CI] update test_ernie_28b_vl.py (#6019)

  Show description for 5afeef6
  

  StareAtYou

  and

  root

  authored
  5afeef6

  Copy full SHA for 5afeef6

• [XPU][CI] Update XVLLM_PATH setup in run_xpu_ci_pytest.sh (#6018)

  Show description for becd8c3
  plusNew001

  authored
  becd8c3

  Copy full SHA for becd8c3

• [BugFix] fix metrics cache tokens (#6001)

  Show description for cb9f952
  kevincheng2

  authored
  cb9f952

  Copy full SHA for cb9f952

• [V1 Loader] Load safetensors weights in natural keyorder (#6006)

  Show description for 8061f74
  

  bukejiyu

  and

  yuanlehome

  authored
  8061f74

  Copy full SHA for 8061f74

• [Optimization] Do not compute ATTN padding part in In Cuda graph mode (#5985)

  zhoutianzi666

  authored
  ad8d05a

  Copy full SHA for ad8d05a

• [BugFix] Fix insert_zmq_task_to_scheduler break bug (#5960)

  Show description for 9c559d0
  ming1753

  authored
  9c559d0

  Copy full SHA for 9c559d0

• [BugFix] Fix entropy calculation issue in TP (#5997)

  Show description for eb8ce36
  Deleter-D

  authored
  eb8ce36

  Copy full SHA for eb8ce36

Commits on Jan 12, 2026

• [Docs] Update FastDeploy version to 2.3.3 in NVIDIA GPU installation documentation (#6010)

  Show description for fe7588d
  

  Copilot

  and

  Jiang-Jia-Jun

  authored
  fe7588d

  Copy full SHA for fe7588d

• [CI] Add fd-router build_task (#5967)

  Show description for 0d3dede
  EmmonsCurse

  authored
  0d3dede

  Copy full SHA for 0d3dede

• [Optimization] Accelerate Qwen3 QK RMSNorm via Fused Triton Kernel (#5880)

  Show description for 2533836
  Sunny-bot1

  authored
  2533836

  Copy full SHA for 2533836

• [ci case]Check the chunking of the chat interface (#5981)

  Show description for 1aa7e82
  

  xjkmfa

  and

  xujing43

  authored
  1aa7e82

  Copy full SHA for 1aa7e82

• [XPU]add ci test cast for P_EP4TP4 D_EP4TP1 (#5988)

  Show description for fefc0b8
  ddchenhao66

  authored
  fefc0b8

  Copy full SHA for fefc0b8

• Support setting communication groups in custom_allreduce and the all-to-all\transpose fused operator during the decoding phase. (#5917)

  carryyu

  authored
  223b2f5

  Copy full SHA for 223b2f5

• [BugFix] [MultiAPIServer] fix rdma script and port check for multi api server (#5935)

  Show description for 60ee72f
  

  liyonghua0910

  and

  Jiang-Jia-Jun

  authored
  60ee72f

  Copy full SHA for 60ee72f

Commits on Jan 10, 2026

• remove decoder_num_blocks_device memset (#5982)

  Sunny-bot1

  authored
  17ef392

  Copy full SHA for 17ef392

• MLA clean code (#5979)

  zhoutianzi666

  authored
  b8d9daa

  Copy full SHA for b8d9daa

Commits on Jan 9, 2026

• dev_fix_mtp_forward_meta (#5976)

  xiaoluomi

  authored
  62bd92f

  Copy full SHA for 62bd92f

• [XPU] fix dp4 (#5946)

  zhupengyang

  authored
  9db48ec

  Copy full SHA for 9db48ec

• [Metax] add ci test file & update run_ci_metax.sh (#5975)

  Show description for 384ffd6
  

  StareAtYou

  and

  root

  authored
  384ffd6

  Copy full SHA for 384ffd6

• [Feature] Support redundant expert for eplb (#5918)

  Show description for 00a01ae
  xiaoxiaohehe001

  authored
  00a01ae

  Copy full SHA for 00a01ae

• [Models] Qwen3VL and Qwen3VL-Moe CUDA graph Support (#5962)

  Show description for e6cdea4
  CSWYF3634076

  authored
  e6cdea4

  Copy full SHA for e6cdea4

Pagination

Previous

Next