onnxruntime/orttraining/orttraining
Wei-Sheng Chin 6160ba0692
Fix aten::_to_copy in DORT (#13682)
`aten::_to_copy` is not exportable to ONNX. In DORT, so it's replaced in 
`_replace_to_copy_with_to`. This replacement logic becomes incorrect in latest PyTorch
commit, and this PR is a fix.

Basically, we examine more key-word attributes passed to
`aten::_to_copy` and if they lead to a type casting operator (i.e.,
mapped to ONNX's Cast), we replace that `aten::_to_copy` with
`aten::to`. Unsupported attributes are removed (with a low risk of
breaking FX graph's assumptions).
2022-11-18 09:31:18 -08:00
..
core Gather to Slice Fusion (#13599) 2022-11-10 13:03:30 +08:00
eager Switch GSL to MS GSL 4.0.0 (#13416) 2022-10-29 04:15:20 -07:00
lazy_tensor [LazyTensor] Make LORT Build Again with Latest PyTorch (#13303) 2022-10-13 13:56:17 -07:00
models Fix typo enviroment => environment (#13195) 2022-10-03 17:02:26 -07:00
python Fix aten::_to_copy in DORT (#13682) 2022-11-18 09:31:18 -08:00
test Fix aten::_to_copy in DORT (#13682) 2022-11-18 09:31:18 -08:00
training_api Fix on-device training ExportModelForInferencing api (#13510) 2022-10-31 21:29:06 -07:00
training_ops [ROCm] Fix bug for rocm ep build using MS GSL 4.0.0 (#13525) 2022-11-01 13:05:55 +08:00