mirror of https://github.com/llvm/torch-mlir
1c2778dd56
Supports asymmetric padding by performing a torch.nn.functional.pad on the input before performing the convolution. Signed-off-by: Suraj Sudhir <suraj.sudhir@arm.com> |
||
|---|---|---|
| .. | ||
| CMakeLists.txt | ||
| DefaultDomainAtoF.cpp | ||
| DefaultDomainGtoP.cpp | ||
| DefaultDomainQtoZ.cpp | ||
| OnnxLstmExpander.cpp | ||
| PassDetail.h | ||
| Passes.cpp | ||
| Patterns.cpp | ||
| TorchOnnxToTorch.cpp | ||
| Utils.cpp | ||