Skip to content

Pull requests: verl-project/verl

New pull request New
435 Open 3,184 Closed
435 Open 3,184 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

[misc] feat: upgrade tensordict to fix nested tensor _ragged_idx problem
#6082 opened Apr 21, 2026 by vermouth1992 Collaborator Loading…
8 tasks
[fix] restore ByteLevel tokenizer for LLaMA-family models under transformers>=5
#6081 opened Apr 20, 2026 by kaixih Contributor Loading…
1
3
[data] Filter Overlonged Prompts in tool_agent_loop
#6079 opened Apr 20, 2026 by HwCARI Contributor Loading…
8 tasks
1
2
[vllm_omni, rollout] fix: forward-compat with vllm-omni multi-LoRA API rename
#6077 opened Apr 20, 2026 by ultranationalism Loading…
6 tasks done
3
[fully_async] fix: use_trainer_do_validate by elastic replica
#6076 opened Apr 20, 2026 by ArronHZG Collaborator Loading…
8 tasks
[worker] Fix: Add missing dump_memory_snapshot in engine workers.
#6075 opened Apr 20, 2026 by Shangwei-Li Contributor Loading…
3 of 8 tasks
4
[veomni] feat: enable VeOmni engine for on-policy distillation
#6072 opened Apr 20, 2026 by hjshi84 • Draft
2
[fully_async] Fix: fix fully async profiler for first step.
#6070 opened Apr 20, 2026 by Shangwei-Li Contributor Loading…
4 of 8 tasks
3
[rollout] feat: enhancing global load balancer (issue #5442)
#6059 opened Apr 19, 2026 by EricWuji Loading…
23
old pr, but previously used to show how to enable fa2. Xpu/fa2 kernel support
#6057 opened Apr 18, 2026 by kahlun Loading…
8 tasks
3
[fully_async, rollout] feat: enable online policy distillation in fully async training
#6056 opened Apr 18, 2026 by xiefan46 • Draft
4 of 8 tasks
3
[rollout] feat: improve error messages for malformed tool calls
#6055 opened Apr 18, 2026 by xiefan46 Loading…
6 of 8 tasks
2
[megatron] fix: correct batch_num_tokens aggregation for context parallelism
#6049 opened Apr 18, 2026 by ianliuy Loading…
5 of 6 tasks
1
5
[rollout] feat: prep for vexact release
#6047 opened Apr 17, 2026 by pengwu22 Collaborator Loading…
4
feat:support exporting fp8 weights in FP8 E2E
#6045 opened Apr 17, 2026 by eternally-z Contributor • Draft
8 tasks
3
[fully_async, reward] feat: enable GenRM/DisRM support in fully async training
#6044 opened Apr 17, 2026 by xiefan46 Loading…
2 of 8 tasks
1
5
[ci] chore: change npu ci machine for test, nomerge
#6043 opened Apr 17, 2026 by yyyy2000 Contributor • Draft
8 tasks
1
[trainer] refactor: code refactor for diffusion training
#6042 opened Apr 17, 2026 by zhtmike Contributor Loading…
6 of 8 tasks
7
[data] fix: pad data in preprocess_packed_seqs if shorter than align_size
#6035 opened Apr 17, 2026 by beirong8kmiles Contributor Loading…
8 tasks
4
[training_utils] fix: Update metrics.py to change calculate_log_prob_diff to calculate_diff
#6033 opened Apr 16, 2026 by victordion Contributor Loading…
2 tasks done
5
[fsdp, sft] feat: add Gemma4 FSDP SFT support
#6030 opened Apr 16, 2026 by khazic Contributor Loading…
2
[trainer, worker] feat: grouped mini-batching and DP balancing for multi-output rollouts in main_ppo_sync
#6025 opened Apr 16, 2026 by guillemgt Contributor Loading…
7 of 8 tasks
7
[ci] fix: engine_mindspeed_llm_rl_job add commit id.
#6020 opened Apr 16, 2026 by pengnuoheng Contributor Loading…
8 tasks
1
[ci] chore: add sglang ci for NPU
#6015 opened Apr 15, 2026 by xiazhahe Contributor Loading…
8 tasks
6
[megatron, ckpt] refactor: optimize mcore ckpt manager impl
#6014 opened Apr 15, 2026 by ETOgaosion Collaborator Loading…
8 tasks
3
ProTip! What’s not been updated in a month: updated:<2026-03-20.