Commits

Branch selector

main

User selector

All users

All time

Commit history

Commits on Jun 5, 2025

• Merge branch 'main' into mine_main

  yzlnew

  committed
  ca9008cView commit details

  Copy full SHA for ca9008c

  Browse repository at this point

• adjust for H20

  yzlnew

  committed
  b0978efView commit details

  Copy full SHA for b0978ef

  Browse repository at this point

• [CI]feat:Add NPU CI action and fallback SFT's e2e test defaults to FSDP1 (#1823)

  Show description for b238297
  

  Crispig

  and

  liaochangyue

  authored
  b238297View commit details

  Copy full SHA for b238297

  Browse repository at this point

• Add DeepSeek 671B GRPO example (#1771)

  Show description for a6f15ae
  yzlnew

  authored
  a6f15aeView commit details

  Copy full SHA for a6f15ae

  Browse repository at this point

• [Mcore] dpskv3 671B (#1694)

  Show description for 2f050a8
  ISEEKYAN

  authored
  2f050a8View commit details

  Copy full SHA for 2f050a8

  Browse repository at this point

• [megatron] moonlight fix per_tensor_generator (#1772)

  Show description for 2858733
  ISEEKYAN

  authored
  2858733View commit details

  Copy full SHA for 2858733

  Browse repository at this point

• [bugfix] Force create checkpoint directory before saving dataloader state. (#1625)

  Show description for 81acbb2
  rj42

  authored
  81acbb2View commit details

  Copy full SHA for 81acbb2

  Browse repository at this point

• [feat] Add support for FSDP2 in GRPO-LoRA (#1844)

  Show description for f7f8b04
  thelongestusernameofall

  authored
  f7f8b04View commit details

  Copy full SHA for f7f8b04

  Browse repository at this point

• add some notes

  yzlnew

  committed
  a326debView commit details

  Copy full SHA for a326deb

  Browse repository at this point

• Add ProRL to README.md (#1855)

  Show description for 2fe47f7
  shizhediao

  authored
  2fe47f7View commit details

  Copy full SHA for 2fe47f7

  Browse repository at this point

• [BugFix][CI] Megatron: add ep CI (#1726)

  Show description for 2a386cf
  ETOgaosion

  authored
  2a386cfView commit details

  Copy full SHA for 2a386cf

  Browse repository at this point

• [refactor] Align name_prefix same behavior for pool and wg (#1851)

  Show description for 5b66489
  hongpeng-guo

  authored
  5b66489View commit details

  Copy full SHA for 5b66489

  Browse repository at this point

• fix batch size validation for Megatron (#1811)

  Show description for 565c496
  yzlnew

  authored
  565c496View commit details

  Copy full SHA for 565c496

  Browse repository at this point

Commits on Jun 4, 2025

• [fix] Adding a default value for `RayWorkerGroup.from_detached(name_prefix=None)` (#1838)

  Show description for 2ed63bb
  hongpeng-guo

  authored
  2ed63bbView commit details

  Copy full SHA for 2ed63bb

  Browse repository at this point

• Add log_generations_to_tensorboard Function (#1841)

  Show description for 5580b0b
  none0663

  authored
  5580b0bView commit details

  Copy full SHA for 5580b0b

  Browse repository at this point

• fix fsdp train save checkpoint bug (#1843)

  Show description for fdf7d51
  

  NileZhou

  and

  zhouyi9

  authored
  fdf7d51View commit details

  Copy full SHA for fdf7d51

  Browse repository at this point

• remove

  yzlnew

  committed
  dc57035View commit details

  Copy full SHA for dc57035

  Browse repository at this point

• Add DeepMath to awesome work list (#1847)

  Show description for 7c49f70
  zwhe99

  authored
  7c49f70View commit details

  Copy full SHA for 7c49f70

  Browse repository at this point

• [sglang] Fix for broadcast_pyobj nccl timeout in sgl rollout with larger model (e.g. 32B) (#1846)

  Show description for 0a5c491
  SwordFaith

  authored
  0a5c491View commit details

  Copy full SHA for 0a5c491

  Browse repository at this point

• add full recompute

  yzlnew

  committed
  98197c0View commit details

  Copy full SHA for 98197c0

  Browse repository at this point

• Merge branch 'main' into mine_main

  yzlnew

  committed
  11b36b7View commit details

  Copy full SHA for 11b36b7

  Browse repository at this point

• fix sglang e2e_sppo test (#1832)

  Show description for fba8f34
  chenhaiq

  authored
  fba8f34View commit details

  Copy full SHA for fba8f34

  Browse repository at this point

• Add LUFFY to awesome work list #1608 (#1816)

  Show description for 996b945
  

  ElliottYan

  and

  eric-haibin-lin

  authored
  996b945View commit details

  Copy full SHA for 996b945

  Browse repository at this point

Commits on Jun 3, 2025

• [logging] misc: update PR template and fix lint (#1806)

  eric-haibin-lin

  authored
  adf775cView commit details

  Copy full SHA for adf775c

  Browse repository at this point

• [sft] trainer: port features (dtype, save_freq, test_freq) from PPO config to SFT config (#1451)

  Show description for 15ca90c
  

  HUGHNew

  and

  eric-haibin-lin

  authored
  15ca90cView commit details

  Copy full SHA for 15ca90c

  Browse repository at this point

• [docs] Add verl-tool in the list of "awesome works using verl" (#1829)

  jdf-prog

  authored
  299dde1View commit details

  Copy full SHA for 299dde1

  Browse repository at this point

• [misc] fix fsdp2 has no _fsdp_wrapped_module in lora collect param (#1822)

  Show description for ed3aec2
  PeterSH6

  authored
  ed3aec2View commit details

  Copy full SHA for ed3aec2

  Browse repository at this point

• [Tracking]feat: support wandb proxy (#1817)

  Show description for de5b2f1
  chenhaiq

  authored
  de5b2f1View commit details

  Copy full SHA for de5b2f1

  Browse repository at this point

• [config] feat: Hardcode moe_router_load_balancing_type to none (#1814)

  Show description for 8540d6c
  vermouth1992

  authored
  8540d6cView commit details

  Copy full SHA for 8540d6c

  Browse repository at this point

• [script] feat: upload qwen3 236b script (#1813)

  Show description for 2fbdcb3
  vermouth1992

  authored
  2fbdcb3View commit details

  Copy full SHA for 2fbdcb3

  Browse repository at this point

• [optimization] feat: move kv cache wakeup after model weights release (#1810)

  Show description for 9910086
  vermouth1992

  authored
  9910086View commit details

  Copy full SHA for 9910086

  Browse repository at this point

• [megatron] fix: critic and reward model load tokenizer from config (#301)

  Show description for 668e5f6
  

  HollowMan6

  and

  ETOgaosion

  authored
  668e5f6View commit details

  Copy full SHA for 668e5f6

  Browse repository at this point

• [dev] fix: note that DP balancing doesn't affect advantage calculation (#1809)

  Show description for 263115c
  tongyx361

  authored
  263115cView commit details

  Copy full SHA for 263115c

  Browse repository at this point

• [recipe] prime: Code example for PRIME (#1714)

  Show description for 7695b8d
  

  ZefanW

  and

  eric-haibin-lin

  authored
  7695b8dView commit details

  Copy full SHA for 7695b8d

  Browse repository at this point

Commits on Jun 2, 2025

• [algo] OPO: add implementations and descriptions for OPO (On-Policy RL with Optimal Reward Baseline) (#1796)

  Show description for a4b1bb7
  

  YRdddream

  and

  eric-haibin-lin

  authored
  a4b1bb7View commit details

  Copy full SHA for a4b1bb7

  Browse repository at this point