master

ascend-robot[master]remove matmul_backward none check

93e97e3a创建于 2025年12月10日历史提交

文件	最后提交记录	最后更新时间
templates	!18176 TORCH MAIN SYNC : compiled_autograd_apply_functional && torch.types.FileLike Merge pull request !18176 from dilililiwhy/main_sync_20250201	1 年前
__init__.py	[master]remove matmul_backward none check Co-authored-by: DaiFu<daifu2@huawei.com> # message auto-generated for no-merge-commit merge: !27733 merge 2512matmul_backward into master [master]remove matmul_backward none check Created-by: daifu1234567 Commit-by: DaiFu Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug **What does this PR do / why do we need it**: 修复matmul_backward中grad传none场景报错问题。 有前反向绑定的api，会在前向中加入unpack（校验none）。问题场景npu跑不通，gpu跑通，是因为npu上多支持了matmul_double_backward，导致matmul_backward上多了unpack校验，把none校验住了，因此需要在生成VariableTypeEverything.cpp的地方对matmul_backward特殊处理，去掉None校验。 **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> Fixes # **Special notes for your reviewers**: 已本地验证： 生成的VariableTypeEverything.cpp中matmul_backward已去掉unpack校验：  改前：  构造的单api脚本本地已跑通：  See merge request: Ascend/pytorch!27733	8 个月前
gen_autograd.py	!12041 Fix api dispatch diff between cpu and npu in compositeimplicitautograd key Merge pull request !12041 from 王广斌/master	2 年前
gen_autograd_functions.py	!9967 Add generate python_functions feature Merge pull request !9967 from 王广斌/master_copy	2 年前
gen_variable_factories.py	!6768 codecheck for not test. Merge pull request !6768 from 王夏夏/master	2 年前
gen_variable_type.py	!24256 npu codegen autograd support symint Merge pull request !24256 from wgb/master	1 年前
utils.py	!12041 Fix api dispatch diff between cpu and npu in compositeimplicitautograd key Merge pull request !12041 from 王广斌/master	2 年前