Skip to content

Pull requests: modelscope/ms-swift

New pull request New
96 Open 3,942 Closed
96 Open 3,942 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(dataset): handle failed non-streaming packing workers
#10152 opened Sep 15, 2026 by Excelius-Wang Contributor Loading…
1 of 4 tasks
fix(train): include all trainable multimodal parameters
#10151 opened Sep 15, 2026 by MrCapricornLiu Contributor Loading…
1 of 4 tasks
fix(reward): count tokens across rollout turns
#10149 opened Sep 15, 2026 by MrCapricornLiu Contributor Loading…
1 of 4 tasks
fix(reward): use normalized ToolBench ROUGE thresholds
#10148 opened Sep 15, 2026 by MrCapricornLiu Contributor Loading…
1 of 4 tasks
fix(reward): preserve nested MathORM expressions
#10147 opened Sep 15, 2026 by MrCapricornLiu Contributor Loading…
1 of 4 tasks
fix(train): choose the correct reverse GaLore projection
#10146 opened Sep 15, 2026 by MrCapricornLiu Contributor Loading…
1 of 4 tasks
fix(train): parse typed Muon optimizer options
#10145 opened Sep 15, 2026 by MrCapricornLiu Contributor Loading…
1 of 4 tasks
fix(dpo): normalize packed loss by preference pair count
#10144 opened Sep 15, 2026 by taking-lying-flat Contributor Loading…
1 task done
fix(train): initialize GaLore through split optimizer hooks
#10143 opened Sep 15, 2026 by MrCapricornLiu Contributor Loading…
1 of 4 tasks
fix(train): check trainable layers separately during checkpointing
#10142 opened Sep 15, 2026 by MrCapricornLiu Contributor Loading…
1 of 4 tasks
1
fix(megatron): balance sampler tails across DP ranks
#10141 opened Sep 15, 2026 by MrCapricornLiu Contributor Loading…
1 of 4 tasks
fix(sampling): preserve dataset columns in Ray dispatch
#10139 opened Sep 15, 2026 by Excelius-Wang Contributor Loading…
1 of 4 tasks
fix(sampling): preserve cached choices during scoring
#10138 opened Sep 15, 2026 by Excelius-Wang Contributor Loading…
1 of 4 tasks
feat(train): log loss for each MRL dimension
#10137 opened Sep 15, 2026 by OysterQAQ Loading…
1 of 4 tasks
fix(infer): propagate streaming generation thread errors
#10133 opened Sep 14, 2026 by Excelius-Wang Contributor Loading…
1 of 4 tasks
1
fix(template): replace rejected responses after the last user message
#10127 opened Sep 14, 2026 by Excelius-Wang Contributor Loading…
fix(utils): make external plugins importable in spawned workers
#10126 opened Sep 14, 2026 by Excelius-Wang Contributor Loading…
1 of 4 tasks
fix(train): disable persistent workers when no workers are used
#10123 opened Sep 14, 2026 by Excelius-Wang Contributor Loading…
feat(megatron): add opt-in chunked full-vocabulary GKD loss
#10114 opened Sep 11, 2026 by taking-lying-flat Contributor Loading…
1 of 4 tasks
surpport deepseek-v4-flash-vision-exp megatron-swift
#10112 opened Sep 11, 2026 by z0o0ey Collaborator Loading…
1 of 4 tasks
[Bugfix][RLHF] fix colocate offload and vLLM wake-up ordering
#10109 opened Sep 11, 2026 by rxlqn Loading…
1 of 4 tasks
1
fix(rlhf): align rollout response prefixes during training
#10107 opened Sep 10, 2026 by taking-lying-flat Contributor Loading…
1 of 4 tasks
1
[bugfix][megatron] align GPU/NPU GKD precision
#10102 opened Sep 10, 2026 by xinyu-0877 • Draft
fix(megatron): normalize GKD loss by global valid-token count
#10096 opened Sep 10, 2026 by wzkk123 • Draft
6 of 10 tasks
[NPU] Migrate Megatron training to Core 0.18 with MegatronAdaptor and TENPU
#10087 opened Sep 9, 2026 by addsubmuldiv Collaborator • Draft
2 of 4 tasks
ProTip! Updated in the last three days: updated:>2026-09-12.