Skip to content

Pull requests: verl-project/verl

New pull request New
330 Open 2,710 Closed
330 Open 2,710 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

build(deps): update transformers requirement from <5.0.0 to <6.0.0 dependencies Pull requests that update a dependency file python Pull requests that update python code
#5293 opened Feb 11, 2026 by dependabot bot Loading…
build(deps): bump sglang[all] from 0.5.2 to 0.5.8.post1 dependencies Pull requests that update a dependency file python Pull requests that update python code
#5292 opened Feb 11, 2026 by dependabot bot Loading…
[misc,trainer,rollout] feat: add Prometheus metrics logging to experiment tracking
#5291 opened Feb 11, 2026 by guillemgt Loading…
7 of 8 tasks
3
Update installation instruction to the actual repo link
#5290 opened Feb 11, 2026 by Godofnothing Loading…
3
[reward] fix: reward model args and reward_kwargs bug
#5289 opened Feb 11, 2026 by yyDing1 Loading…
8 tasks
2
[doc] fix: update recipe link to fix 404 not found
#5286 opened Feb 11, 2026 by tardis-key Loading…
3 of 8 tasks
9
[doc] chore: gspo update config and add version with npu
#5279 opened Feb 11, 2026 by chengminhua Loading…
1 of 8 tasks
[vllm] feat: add model saving func to dump vllm state_dict before and after update weights
#5278 opened Feb 11, 2026 by RobotGF Loading…
8 tasks
2
[rollout] fix: prompt2text decoding for SingleTurnAgentLoop
#5277 opened Feb 11, 2026 by jnash10 Loading…
1
1
[veomni] feat: Add GRPO training scripts for Qwen3-VL-30B-MOE (VeOmni Backends)
#5275 opened Feb 11, 2026 by phdddd Loading…
8 tasks
5
[rollout] fix: use the same request_ids in each rollout turn for better tracking in rollout backend
#5271 opened Feb 10, 2026 by PeterSH6 Loading…
8 tasks
5
[fsdp, vllm] feat: add NPU GRPO training scripts for Qwen3-VL-30B (FSDP/VLLM backends)
#5260 opened Feb 10, 2026 by alwaysyiyu Loading…
8 tasks
7
[algo] feat: add NPU SAPO training script for Qwen3-8B (FSDP/vLLM backends)
#5257 opened Feb 10, 2026 by Vvictorrrr Loading…
2 of 7 tasks
5
[megatron, vllm] feat: NVFP4 (W4A16) QAT training support via ModelOpt
#5254 opened Feb 9, 2026 by jQizhang • Draft
2 of 8 tasks
4
adding CFPO to verl
#5252 opened Feb 9, 2026 by asparius Loading…
4
[megatron] fix: add megatron checkpoint patch
#5251 opened Feb 9, 2026 by Begunner Loading…
8 tasks
5
[fsdp, vllm] feat: add NPU GRPO training scripts for Qwen3-VL-8B (FSDP/VLLM backends)
#5250 opened Feb 9, 2026 by zhihaofang1017 Loading…
8 tasks
4
[fsdp] feat: upcast MoE routing to FP32 for better accuracy
#5249 opened Feb 9, 2026 by Shangwei-Li Loading…
3 of 8 tasks
3
[tool] feature: scheduling analysis based on profiling data
#5248 opened Feb 9, 2026 by mayunaise • Draft
5 of 8 tasks
12
fix: little Ascend NPU implementation bug fix
#5246 opened Feb 9, 2026 by SolarWindRider Loading…
[megatron] feat: add script for qwen3_235b_grpo training on npu platform
#5242 opened Feb 9, 2026 by wangshuyang31 • Draft
8 tasks
4
[rollout,vllm] Fix DP args and local_rank for Ray NOSET_VISIBLE_DEVICES
#5233 opened Feb 7, 2026 by JohnConnor123 Loading…
2 of 3 tasks
3
fix(agent_loop): handle batch size smaller than num_workers
#5231 opened Feb 7, 2026 by aoshen524 Loading…
3 tasks
3
feat(vision): add Vision DP for parallel ViT computation across SP ranks
#5230 opened Feb 7, 2026 by aoshen524 Loading…
3 of 4 tasks
4
[trainer] feat: add per-round logprob mismatch metrics for multi-turn training
#5229 opened Feb 7, 2026 by aoshen524 Loading…
5 of 7 tasks
2
ProTip! Updated in the last three days: updated:>2026-02-08.