Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on May 14, 2026

• [model, fsdp] fix: honor SP-rolled labels in fused kernels (#6068) (#6268)

  Show description for 575d5a8
  shivam2199

  authored
  575d5a8

  Copy full SHA for 575d5a8

Commits on May 13, 2026

• [fsdp, fully_async] fix: fix CI import fast_pos_embed_interpolate in Qwen3-VL (#6332)

  Show description for d98a414
  zhihaofang1017

  authored
  d98a414

  Copy full SHA for d98a414

• [fsdp, fully_async] feat: add Qwen3-VL-30B-A3B fully async GRPO training script on geo3k (#6131)

  Show description for 879ecf9
  zhihaofang1017

  authored
  879ecf9

  Copy full SHA for 879ecf9

• [model, fsdp] fix: Fix modeling_qwen2_5_vl missing attribute 'Qwen2RMSNorm' (#5901)

  Show description for c85e600
  

  ZhuYajun-AI

  and

  gemini-code-assist[bot]

  authored
  c85e600

  Copy full SHA for c85e600

Commits on May 12, 2026

• [megatron] fix: fix bug with mcore0.12.1 + torch2.9.0 (#6322)

  Show description for c311eaa
  yyyy2000

  authored
  c311eaa

  Copy full SHA for c311eaa

Commits on May 9, 2026

• [megatron] fix: fix bugs when using position_ids in cp (#6267)

  Show description for b916f60
  

  3 peopleauthored
  b916f60

  Copy full SHA for b916f60

Commits on May 6, 2026

• [megatron] fix: fix seq_len pad len, and adapt to new mtp_loss api (for megatron dev brance) (#6206)

  Show description for c20d504
  

  zpltys

  and

  zhangshuai122

  authored
  c20d504

  Copy full SHA for c20d504

Commits on Apr 29, 2026

• [BREAKING][misc] refactor: migrate Diffusion RL stack to verl-omni (#6200)

  Show description for a512e90
  SamitHuang

  authored
  a512e90

  Copy full SHA for a512e90

Commits on Apr 24, 2026

• [model] feat: support qwen35 mtp sft/rl (#5898)

  Show description for ed913d1
  zpltys

  authored
  ed913d1

  Copy full SHA for ed913d1

Commits on Apr 19, 2026

• [misc] fix: project name refactor - volcengine -> verl (#6053)

  Show description for 58a361a
  

  ETOgaosion

  and

  ShivenSiwach

  authored
  58a361a

  Copy full SHA for 58a361a

Commits on Apr 16, 2026

• [data] fix: pad data in preprocess_packed_seqs if shorter than align_size (#6001)

  Show description for 45c0f58
  

  beirong8kmiles

  and

  wuxibin89

  authored
  45c0f58

  Copy full SHA for 45c0f58

Commits on Apr 15, 2026

• [megatron] fix: update patch for MLA flashattn forward (#6005)

  Show description for b9d71f9
  HollowMan6

  authored
  b9d71f9

  Copy full SHA for b9d71f9

Commits on Apr 10, 2026

• [megatron] fix: Adjust the attention mask shape for VLM with Megatron on NPU (#5945)

  Show description for 204b3e3
  ZLiao097

  authored
  204b3e3

  Copy full SHA for 204b3e3

Commits on Apr 9, 2026

• Revert "[megatron] fix: Adjust the attention mask shape for VLM with Megatron on NPU" (#5942)

  Show description for 2be1e25
  wuxibin89

  authored
  2be1e25

  Copy full SHA for 2be1e25

Commits on Apr 8, 2026

• [megatron] fix: Adjust the attention mask shape for VLM with Megatron on NPU (#5904)

  Show description for e997236
  ZLiao097

  authored
  e997236

  Copy full SHA for e997236

Commits on Apr 7, 2026

• [trainer] feat: add mindspeedllm backend engine support on NPU. (#5680)

  Show description for 52bf6ab
  

  3 peopleauthored
  52bf6ab

  Copy full SHA for 52bf6ab

• [model] fix: replace inplace += with out-of-place addition in dummy visual forward (#5881)

  Show description for f92bc5c
  reonokiy

  authored
  f92bc5c

  Copy full SHA for f92bc5c

Commits on Apr 3, 2026

• [4/n][trainer] feat: flowgrpo - add diffusers + fsdp engine support (#5802)

  Show description for d5c5daa
  zhtmike

  authored
  d5c5daa

  Copy full SHA for d5c5daa

• [megatron] fix: support critic model (#5870)

  Show description for 1625391
  wuxibin89

  authored
  1625391

  Copy full SHA for 1625391

• [docker, ci] fix: all CIs, transformers upgrade to 5.3.0 and vllm==0.18.0 (#5724)

  Show description for 9275edf
  

  4 peopleauthored
  9275edf

  Copy full SHA for 9275edf

Commits on Apr 1, 2026

• [megatron] feat: support cp for bshd format (#5826)

  Show description for fe1abd8
  wuxibin89

  authored
  fe1abd8

  Copy full SHA for fe1abd8

Commits on Mar 31, 2026

• [megatron] feat: support dynamic CP (#5057)

  Show description for 7e9a07c
  ISEEKYAN

  authored
  7e9a07c

  Copy full SHA for 7e9a07c

Commits on Mar 30, 2026

• [fsdp, model] feat: add qwen3.5 fsdp grpo training support. (#5682)

  Show description for 4045d67
  Zhang1Sheng

  authored
  4045d67

  Copy full SHA for 4045d67

Commits on Mar 28, 2026

• [trainer] feat: support use_remove_padding=False for mindspeed backend (#5768)

  Show description for 57021e3
  ji-huazhong

  authored
  57021e3

  Copy full SHA for 57021e3

Commits on Mar 24, 2026

• [ci] fix: fix various ci failure (#5717)

  Show description for a0c2501
  wuxibin89

  authored
  a0c2501

  Copy full SHA for a0c2501

Commits on Mar 23, 2026

• [megatron] fix: remove llama and qwen2 files (#5707)

  Show description for f056425
  ChengQianqian

  authored
  f056425

  Copy full SHA for f056425

Commits on Mar 21, 2026

• [fsdp,megatron,vllm,trainer,algo] feat: On-Policy Distillation (#5041)

  Show description for 455e44c
  

  JacobHelwig

  and

  wuxibin89

  authored
  455e44c

  Copy full SHA for 455e44c

Commits on Mar 20, 2026

• [megatron] fix: add megatron recomputation backward patch (#5251)

  Show description for 04df110
  Begunner

  authored
  04df110

  Copy full SHA for 04df110

Commits on Mar 13, 2026

• [megatron] feat: model engine support mtp (#5561)

  Show description for 5d73af6
  ArronHZG

  authored
  5d73af6

  Copy full SHA for 5d73af6

• [megatron, model] feat: qwen3.5 example (#5381)

  Show description for ef072ac
  

  ISEEKYAN

  and

  wuxibin89

  authored
  ef072ac

  Copy full SHA for ef072ac

Commits on Mar 8, 2026

• [doc] chore: npu docker support vllm013 (#5471)

  Show description for 6dc2d23
  yyyy2000

  authored
  6dc2d23

  Copy full SHA for 6dc2d23

Commits on Mar 4, 2026

• [megatron] fix: add FP8 block quantization padding for EngineWorker (#5440)

  Show description for c618563
  zpqiu

  authored
  c618563

  Copy full SHA for c618563

Commits on Mar 3, 2026

• [megatron] fix: patch support newer mcore version (#5372)

  Show description for 0a902a9
  HollowMan6

  authored
  0a902a9

  Copy full SHA for 0a902a9

Commits on Feb 27, 2026

• [trainer] feat: add padding for tensor alignment in preprocess_thd_no_padding function (#5410)

  Show description for 32705dc
  RobotGF

  authored
  32705dc

  Copy full SHA for 32705dc

Commits on Feb 25, 2026

• [doc, worker] feat: Enable Megatron-Bridge for MTP (#5323)

  Show description for ea042c2
  HollowMan6

  authored
  ea042c2

  Copy full SHA for ea042c2