diff options
| author | Hanhan Wang <hanchung@google.com> | 2021-03-03 11:46:50 -0800 |
|---|---|---|
| committer | Hanhan Wang <hanchung@google.com> | 2021-03-03 11:47:02 -0800 |
| commit | 83c56aa4ee82acb63327f96f502c4778d972b412 (patch) | |
| tree | 0177a62f1ea1d54a65b758b25b3a1820776bd1d5 /llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp | |
| parent | 8c3a70a78f7e15d1c8f474a9e5316f1d3b5a71d0 (diff) | |
| download | llvm-83c56aa4ee82acb63327f96f502c4778d972b412.zip llvm-83c56aa4ee82acb63327f96f502c4778d972b412.tar.gz llvm-83c56aa4ee82acb63327f96f502c4778d972b412.tar.bz2 | |
[mlir][linalg] Add depthwise_conv_2d_input_nhwc_filter_hwcf to Linalg TC ops.
Different from the definition in Tensorflow and TOSA, the output is [N,H,W,C,M]. This can make transforms easier in LinAlg because the indexing maps are plain. E.g., to determine if the fill op has dependency between the depthwise conv op, the current pipeline only recognizes the dep if they are all projected affine map.
Reviewed By: asaadaldien
Differential Revision: https://reviews.llvm.org/D97798
Diffstat (limited to 'llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp')
0 files changed, 0 insertions, 0 deletions