-
Notifications
You must be signed in to change notification settings - Fork 1.2k
Pull requests: modelscope/ms-swift
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix: pass correct image_patch_size for Qwen3-Omni in fetch_image/fetch_video
#8227
opened Mar 6, 2026 by
xxddccaa
Loading…
[Megatron]Enable CP > 1 for Qwen3.5 and fix missing Megatron logging
#8224
opened Mar 6, 2026 by
yangbofun
Loading…
2 of 4 tasks
[WIP] Moe kernel for qwen3 omni in ascend
#8214
opened Mar 5, 2026 by
jiaqiw09
Loading…
1 of 4 tasks
[megatron]feat: Add routing replay support for Megatron-Swift GRPO
#8196
opened Mar 4, 2026 by
XianlongLi
Loading…
1 of 4 tasks
fix(megatron): mask out padding positions (labels==-100) in MTP loss
#8192
opened Mar 4, 2026 by
ChaosCodes
Loading…
feat: support resume_from_checkpoint=True to auto-find last checkpoint
#8050
opened Feb 14, 2026 by
zhichenggeng
Loading…
1 of 4 tasks
[feat] support frames packing for minicpmv4_5 video processing
#8046
opened Feb 13, 2026 by
fanqiNO1
Loading…
2 of 4 tasks
Add QAT (Quantization-Aware Training) Support Callback
#8042
opened Feb 12, 2026 by
y2logic
Loading…
1 task done
[fix] Pass all fsdp_config values to accelerate via environment variables…
#7962
opened Feb 2, 2026 by
tzteyang
Loading…
1 of 4 tasks
[feat] Support ProFit: Extend DFT with Probability Threshold-based Token Filtering
#7921
opened Jan 28, 2026 by
maybefunctionname
Loading…
1 of 4 tasks
feat: add greedy packing, MiniCPM packing support, and dataset progress tracking
#7904
opened Jan 26, 2026 by
Lollipop
Loading…
fix(megatron): disable checkpointing when calculate KL
#7828
opened Jan 20, 2026 by
zzc0430
Loading…
1 of 4 tasks
[template] Support HunyuanMT1.5-1.8B and HunyuanMT1.5-7B templates
#7351
opened Jan 10, 2026 by
rinne1998
Loading…
support cce、tiledmlp、activation cpu offload
#7169
opened Dec 23, 2025 by
meichangsu1
Loading…
1 of 4 tasks
Previous Next
ProTip!
Mix and match filters to narrow down what you’re looking for.