-
Notifications
You must be signed in to change notification settings - Fork 1.6k
Pull requests: modelscope/ms-swift
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Fix GKD Liger teacher routing
#9841
opened Aug 4, 2026 by
taking-lying-flat
Contributor
Loading…
1 of 4 tasks
Fix provider message preprocessing regressions
#9840
opened Aug 4, 2026 by
taking-lying-flat
Contributor
•
Draft
Guard GRPO against non-finite Liger updates
#9839
opened Aug 3, 2026 by
taking-lying-flat
Contributor
•
Draft
Fix Qwen3.5 multimodal packing kwargs compatibility
#9838
opened Aug 3, 2026 by
taking-lying-flat
Contributor
•
Draft
Support ds v4 flash 0731
#9834
opened Aug 3, 2026 by
tastelikefeet
Collaborator
Loading…
1 of 4 tasks
docs: note OpenAI client base_url for multi-model gateways
#9833
opened Aug 3, 2026 by
seven7763
Loading…
Support model initialization without loading pretrained weights
#9821
opened Jul 30, 2026 by
LiXinYuECNU
Loading…
feat(sft):Add optional fused linear ce for qwen2/qwen3 on npu
#9784
opened Jul 22, 2026 by
yangguang-zhang
Loading…
1 of 4 tasks
feat: support epoch-based checkpoint intervals
#9782
opened Jul 22, 2026 by
tongchen126
Contributor
Loading…
2 of 4 tasks
Fix TRL optional dependency checks with Transformers 5
#9770
opened Jul 20, 2026 by
GodRayyy
Loading…
1 of 4 tasks
fix(dataset): raise a clear error on an empty dataset entry
#9756
opened Jul 16, 2026 by
he-yufeng
Contributor
Loading…
Add optional fast lora path for qwen2/qwen3 on npu
#9754
opened Jul 16, 2026 by
burm95
Loading…
1 of 4 tasks
fix(dataset): raise a clear error when custom dataset_info is not a list
#9751
opened Jul 15, 2026 by
he-yufeng
Contributor
Loading…
fix(security): block SSRF via multimodal media URLs in load_file
#9741
opened Jul 13, 2026 by
he-yufeng
Contributor
Loading…
fix: Call destroy_process_group() at the end of the trainer pipelines
#9726
opened Jul 10, 2026 by
gusario
Contributor
Loading…
1 of 4 tasks
fix(tuners/lora): re-apply lora_dtype after peft _cast_adapter_dtype (#9694)
#9699
opened Jul 7, 2026 by
Anai-Guo
Contributor
Loading…
Fix: MiniCPM-V 4.6 training hangs on text-only samples with DeepSpeed
#9639
opened Jun 24, 2026 by
randydl
Contributor
Loading…
[bugfix] guard _repair_ms_bench against an empty messages list
#9608
opened Jun 20, 2026 by
he-yufeng
Contributor
Loading…
[bugfix] guard load_audio against bytes input
#9606
opened Jun 20, 2026 by
he-yufeng
Contributor
Loading…
1 of 4 tasks
Update gkd_loss.py teacher_prompt supports two formats:
#9562
opened Jun 15, 2026 by
yhy19
Loading…
1 of 4 tasks
[bugfix] add optional MEDIA_DECODE_TIMEOUT to prevent silent media-decode hang
#9541
opened Jun 11, 2026 by
HaozheZhang6
Contributor
Loading…
1 of 4 tasks
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.