Skip to content

Pull requests: modelscope/ms-swift

New pull request New
95 Open 3,763 Closed
95 Open 3,763 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

fix: handle tool messages in last round loss scaling
#9923 opened Aug 15, 2026 by Bakuma-sea Loading…
1 of 4 tasks
[model] deepseek-v4-pro-0813
#9922 opened Aug 15, 2026 by hjh0119 Collaborator Loading…
[Megatron] Fix duplicated GRPO KL gathering
#9921 opened Aug 15, 2026 by taking-lying-flat Contributor Loading…
1
Add Ascend NPU FSDP2 LoRA example for Qwen3.8-27B
#9920 opened Aug 14, 2026 by zheliuyu Loading…
6 of 9 tasks
fix: carry rounded seconds in format_time so it never prints '60s'
#9919 opened Aug 14, 2026 by AlejandroCoronadoN Loading…
Fix/qwen35 cache layout compatibility
#9915 opened Aug 14, 2026 by Ruihan11 Contributor • Draft
1 of 4 tasks
docs: fix Reward Model links in RLHF guides
#9910 opened Aug 14, 2026 by tutao0123 • Draft
1 of 4 tasks
Fix BatchSamplerShard tail sampling
#9907 opened Aug 14, 2026 by taking-lying-flat Contributor Loading…
Fix padding-free single-token sequence boundaries
#9904 opened Aug 13, 2026 by taking-lying-flat Contributor Loading…
1
[Train] Reduce padding-free embedding output memory
#9893 opened Aug 11, 2026 by taking-lying-flat Contributor Loading…
1
test(template): cover MiMo audio placeholders
#9891 opened Aug 11, 2026 by RerankerGuo Contributor Loading…
1 of 4 tasks
5
fix(deploy): handle reasoning-only stream chunks
#9881 opened Aug 11, 2026 by taking-lying-flat Contributor Loading…
1 of 4 tasks
fix(grpo): make revert_runtime_names_to_checkpoint idempotent to avoid double-expanding prefix renamings during vLLM weight sync
#9877 opened Aug 10, 2026 by loveleaves Loading…
5 of 8 tasks
@hjh0119
5
Fix IterablePackingDataset workers to use spawn
#9876 opened Aug 9, 2026 by zupengwang Loading…
1 of 4 tasks
1
Optimize GRPO old policy logprob computation
#9874 opened Aug 8, 2026 by taking-lying-flat Contributor • Draft
fix ppo callbacks
#9872 opened Aug 7, 2026 by leuitong Loading…
3 tasks
2
Fix local/unfused Megatron attention and LoRA on Ascend
#9864 opened Aug 6, 2026 by hazelduan Contributor • Draft
fix(ci): keep the transformers<5.0 pin through the package install
#9862 opened Aug 5, 2026 by he-yufeng Contributor Loading…
1
[Feature] Add --streaming_shard: split the streaming dataset across data-parallel ranks
#9860 opened Aug 5, 2026 by Rapisurazurite Loading…
1 of 4 tasks
fix(sampling): preserve remainder samples across shards
#9859 opened Aug 5, 2026 by RerankerGuo Contributor Loading…
1 of 4 tasks
1
fix(internvl): pass max patch count to vllm
#9850 opened Aug 4, 2026 by RerankerGuo Contributor Loading…
1 of 4 tasks
1
2
fix(infer): align adapter names with batch size
#9847 opened Aug 4, 2026 by RerankerGuo Contributor Loading…
1 of 4 tasks
1
1
Fix provider message preprocessing regressions
#9840 opened Aug 4, 2026 by taking-lying-flat Contributor Loading…
1
Fix Qwen3.5 multimodal packing kwargs compatibility
#9838 opened Aug 3, 2026 by taking-lying-flat Contributor Loading…
1
Support model initialization without loading pretrained weights
#9821 opened Jul 30, 2026 by LiXinYuECNU Loading…
2
ProTip! no:milestone will show everything without a milestone.