master

ascend-robotchore(fsdp2): develop longcat-flash-lite model in fsdp2

518f5f29创建于 12 小时前历史提交

文件	最后提交记录	最后更新时间
__init__.py	[pytorch][feature] optimizer, scheduler and args parser module support in fsdp2 backend Co-authored-by: cjy840282<chenjingyi9@huawei.com> # message auto-generated for no-merge-commit merge: !4044 merge fsdp2 into master [pytorch][feature] optimizer, scheduler and args parser module support in fsdp2 backend Created-by: cjy840282 Commit-by: cjy840282 Merged-by: ascend-robot Description: optimizer, scheduler and args parser module support in fsdp2 backend 1、Optimizer factory for unified construction of single optimizer and EP+FSDP2 multi-optimizer, support Adamw now. 2、Learning rate scheduler factory for unified construction of single scheduler and multi-scheduler, support constant, linear and cosine. 3、Parses the argument class using the command line inputs and yaml configuration. See merge request: Ascend/MindSpeed-LLM!4044	4 个月前
muon.py	style(pytorch): SCA compliance rectification Co-authored-by: zhyebin01<zhangyebin@h-partners.com> # message auto-generated for no-merge-commit merge: !4355 merge master into master style(pytorch): SCA compliance rectification Created-by: zhyebin01 Commit-by: zhyebin01 Merged-by: ascend-robot Description: ## What this PR does / why we need it? SCA compliance rectification ## Does this PR introduce any user-facing change? no ## How was this patch tested? pipeline test passed See merge request: Ascend/MindSpeed-LLM!4355	2 个月前
optimizer.py	chore(fsdp2): develop longcat-flash-lite model in fsdp2 Co-authored-by: guihaowen666<guihaowen@huawei.com> # message auto-generated for no-merge-commit merge: !4344 merge br_master_longcat_flash_lite_fsdp2 into master chore(fsdp2): develop longcat-flash-lite model in fsdp2 Created-by: guihaowen666 Commit-by: guihaowen666 Merged-by: ascend-robot Description: ## What this PR does / why we need it? develop longcat-flash-lite model in fsdp2 ## Does this PR introduce any user-facing change? new model development, no user-facing change ## How was this patch tested? Run the inference task and check whether the model can perform normal dialogs. See merge request: Ascend/MindSpeed-LLM!4344	12 小时前
scheduler.py	[pytorch][bugfix]logger bug fix in scheduler for fsdp2 backend Co-authored-by: cjy840282<chenjingyi9@huawei.com> # message auto-generated for no-merge-commit merge: !4064 merge master into master [pytorch][bugfix]logger bug fix in scheduler for fsdp2 backend Created-by: cjy840282 Commit-by: cjy840282 Merged-by: ascend-robot Description: logger bug fix See merge request: Ascend/MindSpeed-LLM!4064	4 个月前