Ai
101 Star 1.4K Fork 947

GVPMindSpore/mindformers

[poc] add weight decay. 可合并
mindspore-cla/yes
pr-check-pass
!7859 niujunhao 1
审查: +3
2025-12-12 12:08
【r1.7.0】SupportMuon optimizer for DeepseekV3 model
mindspore-cla/yes
stat/needs-squash
!7858 JavaZero 1
审查:
2025-12-12 10:41
同步master 并设置use_legacy_format=False 存在冲突
mindspore-cla/yes
ci-pipeline-failed
stat/needs-squash
!7850 Xinrui Chen 8
审查: +3
2025-12-10 11:27
【docs】doc fix: add kwargs in MFLossMonitor 可合并
mindspore-cla/yes
ci-pipeline-passed
SC-SUCC
pr-check-pass
ai-reviewed
!7849 husichao 12
审查: +3
2025-12-10 11:22
增加deepseek-v3 readme文档,预训练,全参和LoRA微调配置 可合并
mindspore-cla/yes
ci-pipeline-passed
SC-SUCC
pr-check-pass
ai-reviewed
!7842 zzzkeke 24
审查: +3
2025-12-09 10:48
fix get_aux_loss_scale function to include mtp_num_layers parameter fo… 可合并
mindspore-cla/yes
ci-pipeline-passed
SC-SUCC
pr-check-pass
ai-reviewed
!7839 JavaZero 26
审查: +3
2025-12-09 09:58
[poc] apply tele yarn. 可合并
mindspore-cla/yes
pr-check-pass
!7821 niujunhao 2
审查: +3
2025-12-05 17:43
【回合r1.7.0】【bugfix】移除onehot重计算
mindspore-cla/yes
ci-pipeline-passed
SC-SUCC
!7814 JavaZero 8
审查:
2025-12-04 21:37
【r1.7.0】SupportMuon optimizer for DeepseekV3 model
mindspore-cla/yes
stat/needs-squash
!7813 JavaZero 2
审查:
2025-12-04 20:30
Python
1
https://gitee.com/mindspore/mindformers.git
git@gitee.com:mindspore/mindformers.git
mindspore
mindformers
mindformers

搜索帮助