Commits

Branch selector

main

User selector

All users

All time

Commit history

Commits on Apr 21, 2025

• fix vllm version in setup.py (#1186)

  Show description for 64672ae
  hiyouga

  authored
  64672aeView commit details

  Copy full SHA for 64672ae

  Browse repository at this point

Commits on Apr 20, 2025

• [dev] fix: instructions about merging from before using ruff (#1180)

  Show description for 103f901
  tongyx361

  authored
  103f901View commit details

  Copy full SHA for 103f901

  Browse repository at this point

• [AMD] docker: Support AMD (ROCMm Kernel) - Support SGLang (#1179)

  Show description for b0e3f13
  yushengsu-thu

  authored
  b0e3f13View commit details

  Copy full SHA for b0e3f13

  Browse repository at this point

• [ray] fix: ray hang due to num_cpus (#1009)

  Show description for 725c676
  tongyx361

  authored
  725c676View commit details

  Copy full SHA for 725c676

  Browse repository at this point

• [CI] fix: add additional pre-commit test before ppo trainer tests (#1175)

  eric-haibin-lin

  authored
  5313d96View commit details

  Copy full SHA for 5313d96

  Browse repository at this point

• [algo] feat: Add DrGRPO (#990)

  Show description for 6d8f2f6
  sunjin-k

  authored
  6d8f2f6View commit details

  Copy full SHA for 6d8f2f6

  Browse repository at this point

• [logging] refactor: use 'from e' for exception stack trace (#1177)

  Show description for b0e2a0a
  YeonwooSung

  authored
  b0e2a0aView commit details

  Copy full SHA for b0e2a0a

  Browse repository at this point

• [Config] fix: disable XFORMERS by default since we immgrated to newer vLLM versions (#1178)

  tongyx361

  authored
  28e45cbView commit details

  Copy full SHA for 28e45cb

  Browse repository at this point

• [megatron] fix: avoid initialization of Megatron if not use (#1143)

  Show description for 3c46da5
  

  BearBiscuit05

  and

  eric-haibin-lin

  authored
  3c46da5View commit details

  Copy full SHA for 3c46da5

  Browse repository at this point

• [megatron] fix optimizer config (#1104)

  ccclyu

  authored
  1ab271eView commit details

  Copy full SHA for 1ab271e

  Browse repository at this point

• [mcore] qwen2moe support (#1139)

  Show description for 4fa7ed6
  ISEEKYAN

  authored
  4fa7ed6View commit details

  Copy full SHA for 4fa7ed6

  Browse repository at this point

• docs: update recent news and logo (#1173)

  eric-haibin-lin

  authored
  c54ec18View commit details

  Copy full SHA for c54ec18

  Browse repository at this point

• Fix ImportError in is_megatron_core_available() and is_vllm_available() Functions (#1131)

  Show description for b39c021
  none0663

  authored
  b39c021View commit details

  Copy full SHA for b39c021

  Browse repository at this point

• docs: add ReTool (#1154)

  eric-haibin-lin

  authored
  0fd56b2View commit details

  Copy full SHA for 0fd56b2

  Browse repository at this point

• change deepspeedai url site. (#1171)

  Show description for f49c531
  

  湛露先生

  authored
  f49c531View commit details

  Copy full SHA for f49c531

  Browse repository at this point

Commits on Apr 19, 2025

• [CI] fix: only check changed files in CI (#1168)

  Show description for 121f0b0
  tongyx361

  authored
  121f0b0View commit details

  Copy full SHA for 121f0b0

  Browse repository at this point

• revert multinode first (#1161)

  Show description for 8719371
  ETOgaosion

  authored
  8719371View commit details

  Copy full SHA for 8719371

  Browse repository at this point

• Update test_dapo_7b.sh to remove extra line (#1155)

  Show description for 6effd52
  BaiqingL

  authored
  6effd52View commit details

  Copy full SHA for 6effd52

  Browse repository at this point

• [BREAKING] ray: rewrite multi-node doc (#1160)

  Show description for f3dc1d7
  ETOgaosion

  authored
  f3dc1d7View commit details

  Copy full SHA for f3dc1d7

  Browse repository at this point

• CI: limit ruff checks and enable push tests (#1157)

  eric-haibin-lin

  authored
  568239fView commit details

  Copy full SHA for 568239f

  Browse repository at this point

Commits on Apr 18, 2025

• [dev] feat: immigrate from yapf & pylint to ruff based on pre-commit (#1010)

  Show description for b00f77d
  tongyx361

  authored
  b00f77dView commit details

  Copy full SHA for b00f77d

  Browse repository at this point

• Doc: add a environment to fix that the memory capacity is unbalanced (#1105)

  Show description for c98fb31
  

  minleminzui

  and

  ocss884

  authored
  c98fb31View commit details

  Copy full SHA for c98fb31

  Browse repository at this point

• [misc] add sglang support for hdfs file load (#1060)

  BearBiscuit05

  authored
  ec59b87View commit details

  Copy full SHA for ec59b87

  Browse repository at this point

• [misc] qwen moe patch for not find 'load_weights' func (#1137)

  BearBiscuit05

  authored
  0bdf7f4View commit details

  Copy full SHA for 0bdf7f4

  Browse repository at this point

Commits on Apr 17, 2025

• [dapo] fix: fix timer for dapo (#1075)

  Show description for ba988bb
  lilei199908

  authored
  ba988bbView commit details

  Copy full SHA for ba988bb

  Browse repository at this point

• Move entropy to comput log probs to reduce peak memory when calculating entropy. (#1100)

  Show description for 25b0f22
  ETOgaosion

  authored
  25b0f22View commit details

  Copy full SHA for 25b0f22

  Browse repository at this point

• [mcore] resharding model weights by per tensor (#1107)

  Show description for 19d0d07
  BearBiscuit05

  authored
  19d0d07View commit details

  Copy full SHA for 19d0d07

  Browse repository at this point

• refactor: main generation should also use pad/unpad from verl.protocol (#1103)

  Show description for 833e7d7
  0x404

  authored
  833e7d7View commit details

  Copy full SHA for 833e7d7

  Browse repository at this point

• fix: loading HF model in rank0 for mcore megatron model (#998)

  Show description for f04a6db
  ann-qin-lu

  authored
  f04a6dbView commit details

  Copy full SHA for f04a6db

  Browse repository at this point

• [mcore] refactor (#1064)

  Show description for be9def6
  ISEEKYAN

  authored
  be9def6View commit details

  Copy full SHA for be9def6

  Browse repository at this point

Commits on Apr 16, 2025

• [sft] feat: Add WSD (Warmup-Stable-Decay) scheduler for SFT (#1041)

  Show description for d6821a0
  

  3 peopleauthored
  d6821a0View commit details

  Copy full SHA for d6821a0

  Browse repository at this point

• [feat] video inputs (#1116)

  Show description for 54fbd15
  mingruimingrui

  authored
  54fbd15View commit details

  Copy full SHA for 54fbd15

  Browse repository at this point

• fix: missed loss_agg_mode in dp_actor (#945)

  Show description for 3e3d937
  dirtyDan0

  authored
  3e3d937View commit details

  Copy full SHA for 3e3d937

  Browse repository at this point

• Update install.rst to fix a typo (#1111)

  Show description for f845a46
  silverriver

  authored
  f845a46View commit details

  Copy full SHA for f845a46

  Browse repository at this point

• [misc] add dummy load for sglang (#1068)

  BearBiscuit05

  authored
  635997aView commit details

  Copy full SHA for 635997a

  Browse repository at this point