Skip to content

Pull requests: NovaSky-AI/SkyRL

New pull request New
184 Open 1,357 Closed
184 Open 1,357 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

Persistent MXFP8 for experts layers in MoE
#1940 opened Jul 23, 2026 by kailash109 • Draft
2
Keep prewarmed inference alive after model unload
#1939 opened Jul 23, 2026 by j316chuck Contributor Loading…
4
FSDP Concurrent Multi-LoRA Training
#1938 opened Jul 22, 2026 by atemaguer Contributor Loading…
7
MXFP8 for experts in MoE
#1937 opened Jul 22, 2026 by kailash109 • Draft
5
[deps] support GB200 machines
#1936 opened Jul 22, 2026 by erictang000 Collaborator Loading…
4
[feat] Add Delta Weight Sync Support in SkyRL
#1935 opened Jul 22, 2026 by SumanthRH Member • Draft
[feat][SFT] Cloud-path (S3/GCS) dataset ingestion for pretokenized and text datasets
#1933 opened Jul 21, 2026 by avigyabb Collaborator • Draft
[logging] Publish training metrics and step/epoch to Prometheus
#1931 opened Jul 20, 2026 by eicherseiji Contributor Loading…
3
[feat] Add suppport for mooncake connector for P/D; custom P/D init kwargs
#1926 opened Jul 17, 2026 by SumanthRH Member • Draft
[logging] Account for env setup and overhead in trajectory time
#1925 opened Jul 17, 2026 by eicherseiji Contributor • Draft
[logging] Log rollout buffer depth to W&B and Prometheus
#1924 opened Jul 17, 2026 by eicherseiji Contributor Loading…
10
[logging] Publish the training loop phase as a Prometheus gauge
#1923 opened Jul 17, 2026 by eicherseiji Contributor Loading…
10
[tinker] Fix per-tenant blocking on forward/forward_backward calls
#1920 opened Jul 17, 2026 by kailash109 Loading…
@erictang000
3
[feat] Adding rollout router replay (R3) on tinker path
#1919 opened Jul 17, 2026 by kailash109 Loading…
@erictang000
8
[GLM5.2] Use official IndexShare support
#1917 opened Jul 17, 2026 by yapdianang • Draft
perf(sample-support): add packed base64 transport
#1915 opened Jul 16, 2026 by dyurk-lila Contributor Loading…
@erictang000
1
feat(fsdp): add bounded sampler-support replay
#1914 opened Jul 16, 2026 by dyurk-lila Contributor Loading…
@erictang000
1
feat(megatron): add bounded sampler-support replay
#1913 opened Jul 16, 2026 by dyurk-lila Contributor Loading…
@erictang000
1
refactor(generators): assemble routed-expert traces incrementally
#1912 opened Jul 16, 2026 by dyurk-lila Contributor Loading…
@erictang000
1
refactor(inference): extract reusable RemoteInferenceGenerator
#1911 opened Jul 16, 2026 by dyurk-lila Contributor Loading…
@erictang000
7
perf(r3): expand replay routes to local layers only
#1910 opened Jul 16, 2026 by dyurk-lila Contributor Loading…
@erictang000
1
perf(r3): accelerate routed-expert transport with packed arrays
#1909 opened Jul 16, 2026 by dyurk-lila Contributor Loading…
@erictang000
1
fix(r3): align routed-expert metadata and padding
#1908 opened Jul 16, 2026 by dyurk-lila Contributor Loading…
@erictang000
1
[fix] Render multimodal sample prompts under the base model (#1860)
#1907 opened Jul 16, 2026 by sunnyadn Loading…
1
2
FP8 KV-Cache Compatibility for Hybrid vLLM Models
#1899 opened Jul 14, 2026 by YJHMITWEB Contributor Loading…
@erictang000 @jinghanyao1-hub
2
ProTip! Type g p on any issue or pull request to go back to the pull request listing page.