-
Notifications
You must be signed in to change notification settings - Fork 351
Pull requests: NovaSky-AI/SkyRL
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[Fix] Fix race condition with memory consumption by using
ray_init_fixture
#1787
opened Jun 14, 2026 by
SumanthRH
Member
Loading…
[train] Save HF processor on checkpoint export for VLMs
#1785
opened Jun 14, 2026 by
dinhxuanvu
Contributor
Loading…
1 of 2 tasks
[logging] Flush pending step metrics to the tracker when training crashes
#1784
opened Jun 13, 2026 by
discobot
Loading…
[fix] Honor served_model_name and surface HTTP errors in RemoteInferenceEngine
#1783
opened Jun 13, 2026 by
discobot
Loading…
[fix] Use masked mean in advantage batch normalization
#1782
opened Jun 12, 2026 by
discobot
Loading…
Support top-K distillation (SDFT/OPSD): teacher top-K sampling + soft-target CE training
#1777
opened Jun 11, 2026 by
atemaguer
Contributor
Loading…
Rollout Routing Replay (R3) for the
fsdp backend, using vllm==0.22
#1772
opened Jun 10, 2026 by
jamesbraza
Contributor
Loading…
Attach the failed actor's log tail when inference engines die during init
#1771
opened Jun 10, 2026 by
jamesbraza
Contributor
Loading…
[WIP] Add seq packing support for qwen3.5
#1769
opened Jun 10, 2026 by
erictang000
Collaborator
•
Draft
[megatron] Fused LM-head log-prob + entropy (avoid full [*, seq, vocab] logit materialization)
#1765
opened Jun 9, 2026 by
dyurk-lila
Loading…
4 tasks done
[train] VLM SFT support on Megatron backend (Qwen3-VL)
#1752
opened Jun 4, 2026 by
s-chundi
Loading…
feat(profiler): drive torch.profiler around the training loop
#1750
opened Jun 4, 2026 by
dyurk-lila
Loading…
fix: release slot on CancelledError instead of hard-exit at epoch boundaries
#1733
opened Jun 1, 2026 by
dinhxuanvu
Contributor
Loading…
add validation sample logging
#1713
opened May 28, 2026 by
erictang000
Collaborator
Loading…
2 tasks
feat: add max_training_steps config for CI smoke tests
#1690
opened May 19, 2026 by
dinhxuanvu
Contributor
Loading…
3 of 4 tasks
[workers] Clean teardown on SIGTERM: drain CUDA + destroy process group
#1688
opened May 18, 2026 by
xyuzh
Loading…
[tx] add retries for getting random port used for JAX coordinator address
#1652
opened May 12, 2026 by
andrewsykim
Contributor
Loading…
Avoid blocking Ray registry lookups for policy loss resolution
#1644
opened May 10, 2026 by
taivu1998
Loading…
Previous Next
ProTip!
Find all pull requests that aren't related to any open issues with -linked:issue.