master

ascend-robot[mindspore]fix distrib_optim save

3a6ca5b9创建于 1月22日历史提交

文件	最后提交记录	最后更新时间
swap_optimizer	!2793 [mindspore][master] add swap-optimizer patch Merge pull request !2793 from wangjialin/swap_optimizer_pin_memory	8 个月前
__init__.py	!2585 megatron v012 adaptor Merge pull request !2585 from 张珂/megatron012_mindspeed	10 个月前
adamw.py	!2789 [mindspore][master]modify npu_apply_fused_adamw_v2 to msadapter Merge pull request !2789 from minghu111/master0828	8 个月前
distrib_optimizer.py	[mindspore]fix distrib_optim save Co-authored-by: ybwang19<1605891897@qq.com> # message auto-generated for no-merge-commit merge: !3218 merge master into master [mindspore]fix distrib_optim save Created-by: ybwang19 Commit-by: ybwang19 Merged-by: ascend-robot Description: [🔗关联问题单](https://e.gitee.com/mind_spore/dashboard?issue=IDJ2KI) 修改前后PT和MS侧断点续训精度不齐：   修改后PT和MS侧断点续训精度对齐：   See merge request: Ascend/MindSpeed!3218	4 个月前
optimizer.py	[mindspore][bugfix] bugfix for chatglm3-6b pretrain Co-authored-by: wanglijun55<wanglijun54@huawei.com> # message auto-generated for no-merge-commit merge: !2996 merge branch-glm3 into master [mindspore][bugfix] bugfix for chatglm3-6b pretrain Created-by: wanglijun55 Commit-by: wanglijun55 Merged-by: ascend-robot Description: **修改说明：** 1、关联问题单：https://e.gitee.com/mind_spore/issues/table?issue=ID5HCI chatglm3-6b pretrain报错syncstream failed，定位原因为原patch中out.untyped_storage().resize_(0)的写法会导致loss tensor未初始化，确认改为out.data = torch.empty([]) 2、由于deallocate_output_tensor_ patch写法变更，custom_backward patch中assert loss.numel() == 1的判断和megatron保持一致，因而删除custom_backward patch **自验证结果：** 1、chatglm3-6b pretrain功能正常  2、qwen3-235b finetune功能正常  See merge request: Ascend/MindSpeed!2996	6 个月前