Commits

Branch selector

main

User selector

All users

All time

Commit history

Commits on Feb 27, 2025

• [feat] tracking support tensorboard (#408)

  Show description for 82b38e2
  iceflame89

  authored
  82b38e2View commit details

  Copy full SHA for 82b38e2

  Browse repository at this point

• [ckpt] fix: fix oom when resume from ckpt (#402)

  vermouth1992

  authored
  a0f05daView commit details

  Copy full SHA for a0f05da

  Browse repository at this point

• [fix] Fix evaluation file path in remax training scripts. (#404)

  Show description for 052b0a3
  yaguanghu

  authored
  052b0a3View commit details

  Copy full SHA for 052b0a3

  Browse repository at this point

• [ckpt] replace DataLoader with StatefulDataLoader to support resume training for SequentialSampler (#389)

  Show description for 96d98cc
  zpqiu

  authored
  96d98ccView commit details

  Copy full SHA for 96d98cc

  Browse repository at this point

• [misc] fix: disable chunked-prefill by default (#259)

  Show description for 558fae5
  PeterSH6

  authored
  558fae5View commit details

  Copy full SHA for 558fae5

  Browse repository at this point

• [ci] fix: fix qwen0.5b megatron ci (#396)

  vermouth1992

  authored
  5964358View commit details

  Copy full SHA for 5964358

  Browse repository at this point

Commits on Feb 26, 2025

• apis: add data proto to documentation page. use copy_to_local instead of copy_local_path_from_hdfs (#358)

  eric-haibin-lin

  authored
  2440aa6View commit details

  Copy full SHA for 2440aa6

  Browse repository at this point

• [misc] add assertion for normalized ppo mini_batch_size and ppo micro… (#382)

  Show description for efd0061
  PeterSH6

  authored
  efd0061View commit details

  Copy full SHA for efd0061

  Browse repository at this point

Commits on Feb 25, 2025

• [fix] Passing ppo_epochs to dp_actor.py (#346)

  Show description for b4c13ce
  jayl940712

  authored
  b4c13ceView commit details

  Copy full SHA for b4c13ce

  Browse repository at this point

• [Fix] Using an enumeration class to avoid spelling errors in adv_esti… (#377)

  Show description for 3b1aef2
  

  thomZ1

  and

  Thom

  authored
  3b1aef2View commit details

  Copy full SHA for 3b1aef2

  Browse repository at this point

• fix spelling error (#374)

  wangchengnuo

  authored
  0dc8e85View commit details

  Copy full SHA for 0dc8e85

  Browse repository at this point

• [ppo] fix: fix minibatch size when n > 1 for megatron worker (#370)

  vermouth1992

  authored
  bbda2e5View commit details

  Copy full SHA for bbda2e5

  Browse repository at this point

Commits on Feb 24, 2025

• rollout: Fix navive_rollout class names. (#361)

  Show description for 656accb
  

  湛露先生

  authored
  656accbView commit details

  Copy full SHA for 656accb

  Browse repository at this point

• [docs] modify the comments (#363)

  BearBiscuit05

  authored
  8eb22b5View commit details

  Copy full SHA for 8eb22b5

  Browse repository at this point

• feat: add support for ulysses sequence parallel for transformers >= 0.48 (#357)

  Show description for d36422b
  xffxff

  authored
  d36422bView commit details

  Copy full SHA for d36422b

  Browse repository at this point

• [fix] Improve the params template for generation (#351)

  Show description for e53dcdb
  BearBiscuit05

  authored
  e53dcdbView commit details

  Copy full SHA for e53dcdb

  Browse repository at this point

• [Fix] Deprecate `val_batch_size` (#353)

  Show description for 4011f40
  tongyx361

  authored
  4011f40View commit details

  Copy full SHA for 4011f40

  Browse repository at this point

Commits on Feb 23, 2025

• feat: tracking support vemlp (#339)

  Show description for 7a128c1
  

  liudayuan-carrot

  and

  liudayuan.carrot

  authored
  7a128c1View commit details

  Copy full SHA for 7a128c1

  Browse repository at this point

• chore: update optimizer_config.py (#348)

  eltociear

  authored
  32a3d62View commit details

  Copy full SHA for 32a3d62

  Browse repository at this point

Commits on Feb 22, 2025

• docs: add links for rloo and volcengine distributed training doc (#343)

  eric-haibin-lin

  authored
  d02b8c7View commit details

  Copy full SHA for d02b8c7

  Browse repository at this point

• algo: Rloo advantage estimator (#341)

  Show description for 6a820b6
  ZefanW

  authored
  6a820b6View commit details

  Copy full SHA for 6a820b6

  Browse repository at this point

Commits on Feb 21, 2025

• docs: add faq for vllm illegal memory access (#333)

  eric-haibin-lin

  authored
  76352aeView commit details

  Copy full SHA for 76352ae

  Browse repository at this point

• [misc] Add Ray Serve to requirements to support multi-node training (#318)

  Show description for 0268917
  uygnef

  authored
  0268917View commit details

  Copy full SHA for 0268917

  Browse repository at this point

Commits on Feb 20, 2025

• distro: bump up version to v0.2.0.dev, limit vllm version (#327)

  eric-haibin-lin

  authored
  0a1b16fView commit details

  Copy full SHA for 0a1b16f

  Browse repository at this point

Commits on Feb 19, 2025

• [megatron] feat: support qwen2 megatron backend (#261)

  Show description for 9448762
  kinman0224

  authored
  9448762View commit details

  Copy full SHA for 9448762

  Browse repository at this point

• fix vllm 0.7 documentation link in readme (#317)

  vermouth1992

  authored
  35c8daeView commit details

  Copy full SHA for 35c8dae

  Browse repository at this point

• fix: specify the hash version of action in scorecard.yml (#313)

  WillemJiang

  authored
  39d175bView commit details

  Copy full SHA for 39d175b

  Browse repository at this point

• docs: add an example for Ray on Slurm (#309)

  Show description for 55a4d3c
  danielz02

  authored
  55a4d3cView commit details

  Copy full SHA for 55a4d3c

  Browse repository at this point

• docs: add recent event and blogs (#305)

  eric-haibin-lin

  authored
  8cc37e1View commit details

  Copy full SHA for 8cc37e1

  Browse repository at this point

• Added the dependabot action (#304)

  WillemJiang

  authored
  6856479View commit details

  Copy full SHA for 6856479

  Browse repository at this point

• Added content permissions of the workflow (#303)

  Show description for dd09d47
  WillemJiang

  authored
  dd09d47View commit details

  Copy full SHA for dd09d47

  Browse repository at this point

Commits on Feb 18, 2025

• fix: fix offload/load optimizer impl (#299)

  Show description for cf54dd1
  Mofan-Z

  authored
  cf54dd1View commit details

  Copy full SHA for cf54dd1

  Browse repository at this point

• docs: update news

  eric-haibin-lin

  authored
  bb939b7View commit details

  Copy full SHA for bb939b7

  Browse repository at this point

• example: fix the gemma2 example, update NGC dockerfile (#291)

  eric-haibin-lin

  authored
  77f065eView commit details

  Copy full SHA for 77f065e

  Browse repository at this point

Commits on Feb 17, 2025

• Fix wrongs args desc . (#294)

  Show description for 0dfcb7f
  

  湛露先生

  authored
  0dfcb7fView commit details

  Copy full SHA for 0dfcb7f

  Browse repository at this point