Skip to content

Pull requests: NovaSky-AI/SkyRL

New pull request New
116 Open 1,032 Closed
116 Open 1,032 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

[tinker] Support prompt_logprobs in SkyRLTrainBackend sample() path
#1461 opened Apr 6, 2026 by pbokc • Draft
6
[tinker] Support KL loss in SkyRLTrainBackend
#1460 opened Apr 5, 2026 by pbokc Loading…
4
[tinker] Forbid extra keys in EngineConfig
#1459 opened Apr 4, 2026 by pbokc Loading…
[WIP] Add OpenReward environment integration example
#1458 opened Apr 4, 2026 by tyfeng1997 • Draft
2 of 4 tasks
1
2
feat: LLM-synthesized hints for failed trajectories
#1456 opened Apr 4, 2026 by dzorlu Loading…
4 tasks
6
[skyrl] New inference client in the skyrl-train tinker backend
#1452 opened Apr 3, 2026 by nithinvc • Draft
1 task
[skyrl-train] feat: add native GMPO policy loss with validation and tests
#1449 opened Apr 2, 2026 by taivu1998 Loading…
3
Fix event-loop blocking in one-step-off async save/export paths
#1446 opened Apr 2, 2026 by taivu1998 Loading…
1
3
Change default KL estimator from k3 to k2 for loss-based KL
#1445 opened Apr 2, 2026 by taivu1998 Loading…
2
[skyrl-train] Flip grpo_norm_by_std default to false
#1443 opened Apr 2, 2026 by taivu1998 Loading…
1
[skyrl-train] Add trainer-side max_response_length for Dr. GRPO normalization and DAPO overlong handling
#1440 opened Apr 2, 2026 by taivu1998 Loading…
2
[R3] Enable R3 with new inference
#1428 opened Apr 2, 2026 by hao-aaron Loading…
15
[megatron] support qwen3.5 models for megatron, bump mbridge + megatron-core to latest
#1425 opened Apr 1, 2026 by erictang000 Loading…
8
[WIP][tx] Add initial implementation of RayJaxBackend
#1418 opened Mar 31, 2026 by andrewsykim • Draft
1
1
[SkyRL][train] Support prompt_logprobs in /sample in the new inference stack
#1417 opened Mar 31, 2026 by nithinvc Loading…
6 tasks done
@SumanthRH
32
Fleet/all
#1410 opened Mar 30, 2026 by dzorlu Loading…
14
[train] Add Virtual Pipeline Parallelism support to Megatron
#1400 opened Mar 27, 2026 by tamoghnokandar Loading…
1
@erictang000
12
[tinker] Support PPO loss with Tinker and add critic model in SkyRLTrainBackend
#1389 opened Mar 25, 2026 by tamoghnokandar Loading…
4 tasks done
@SumanthRH
12
log all trajectory repetitions per training step (not just first)
#1388 opened Mar 25, 2026 by arteemg Loading…
5
[CI] Migrate MoE Tests to OlMoE
#1384 opened Mar 25, 2026 by devpatelio • Draft
2
[train] Prefix-aware merge for step-wise trajectories (#1277)
#1377 opened Mar 24, 2026 by deepsheth3 Loading…
9
Ulysses position_ids pre-gather, NUMA rewrite, and operational improvements
#1371 opened Mar 23, 2026 by ashutoshuiuc Loading…
@SumanthRH
9
FSDP2/MoE core training stability fixes
#1370 opened Mar 23, 2026 by ashutoshuiuc Loading…
@SumanthRH
27
Add Anthropic Messages API (/v1/messages) endpoint
#1369 opened Mar 23, 2026 by ashutoshuiuc Loading…
1
16
Improved LoRA weight swap and robust transitions_to_training_data
#1368 opened Mar 23, 2026 by ashutoshuiuc Loading…
1
@erictang000
10
ProTip! Exclude everything labeled bug with -label:bug.