Skip to content

Pull requests: NovaSky-AI/SkyRL

New pull request New
100 Open 964 Closed
100 Open 964 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

[tx] Fuse the projection matrices for Qwen 3.5 tx
#1362 opened Mar 21, 2026 by pcmoritz Loading…
2
[skyrl] Add ImageChunk and ImageAssetPointerChunk types
#1361 opened Mar 21, 2026 by nithinvc Loading…
2 tasks done
13
[train] LoRA-only weight syncing for Megatron
#1360 opened Mar 20, 2026 by hao-aaron Loading…
1
@erictang000
17
feat: add multi-node WandB system metrics aggregation
#1358 opened Mar 20, 2026 by penfever Loading…
3 tasks
7
feat: make max_logprobs configurable for inference engines
#1357 opened Mar 20, 2026 by penfever Loading…
2 tasks
fix: stagger vLLM engine startup to avoid EADDRINUSE
#1356 opened Mar 20, 2026 by penfever Loading…
2 tasks
3
[train] Enable support for the mp backend with the new inference codepath
#1355 opened Mar 20, 2026 by SumanthRH • Draft
Fix/causal conv1d cuda128
#1354 opened Mar 20, 2026 by dzorlu Loading…
8
[bug][examples] Fix lcb dataset issue in examples dir
#1349 opened Mar 19, 2026 by s-chundi Loading…
1
5
[train][2/N] Support for Megatron PP + CP for R3
#1335 opened Mar 17, 2026 by devpatelio Loading…
10
[tx] Enable MPS backend for Apple Silicon on Jax backend tx
#1332 opened Mar 17, 2026 by pcmoritz Loading…
3 tasks
2
[train] Multi-modal inputs support in FSDP2
#1331 opened Mar 17, 2026 by nithinvc Loading…
@SumanthRH
28
Glm r3
#1326 opened Mar 14, 2026 by erictang000 • Draft
WIP Qwen 3.5 Dense FSDP
#1313 opened Mar 11, 2026 by CharlieFRuan • Draft
[bug][train] Fix max_seq_len calculation
#1303 opened Mar 10, 2026 by tamoghnokandar Loading…
1
@SumanthRH
7
MoE/FSDP2 training fixes, Anthropic Messages API, and robustness improvements
#1298 opened Mar 9, 2026 by ashutoshuiuc Loading…
4 tasks done
2
26
[skyrl-train] Implement loss reduction via advantage normalization and fix token_mean reduction strategy
#1296 opened Mar 9, 2026 by justinvyu Loading…
6
[train][FullyAsync] Implement customizable weight sync frequency
#1293 opened Mar 8, 2026 by tamoghnokandar Loading…
@SumanthRH @CharlieFRuan
4
feat(train): prefix-aware merge for step-wise trajectories (#1277)
#1289 opened Mar 6, 2026 by deepsheth3 Loading…
@CharlieFRuan
11
[tinker][SkyRL] Add sample and renderer API to new inference client
#1287 opened Mar 6, 2026 by nithinvc • Draft
39
WIP: return_dict=False fixes + H200 validation scripts
#1280 opened Mar 5, 2026 by tyler-griggs • Draft
[train] Add sampler-as-reference KL option
#1263 opened Mar 3, 2026 by tyler-griggs • Draft
2 tasks
[train] Add KL-in-advantages mode
#1262 opened Mar 3, 2026 by tyler-griggs • Draft
2 tasks
[train] Add importance weight diagnostics and fix IS loss overflow
#1261 opened Mar 3, 2026 by tyler-griggs • Draft
2 tasks
[train] Add "sum" loss reduction mode
#1260 opened Mar 3, 2026 by tyler-griggs • Draft
1 task
ProTip! Exclude everything labeled bug with -label:bug.