mirror of https://github.com/llvm/torch-mlir
1be604bfd3
This commit lowers `aten.matmul` to `linalg.BatchMatmul` under the following conditions: 1. The result of matrix multiplication must have batch dimensions, i.e., rank greater than 2. 2. The resultant matrix must have at most 1 dynamic batch dimension. It also handles broadcasting of batch dimensions when batch dimensions of the matrices are broadcastable. Signed-Off-by: Gaurav Shukla <gaurav@nod-labs.com> |
||
|---|---|---|
| .. | ||
| TorchToLinalg | ||
| TorchToSCF | ||
| TorchToStd | ||
| TorchToTMTensor | ||
| TorchToTosa | ||
| Utils | ||
| CMakeLists.txt | ||
| PassDetail.h | ||
| Passes.cpp | ||