diff options
| author | Felix Schneider <fx.schn@gmail.com> | 2023-12-01 22:16:00 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2023-12-01 22:16:00 +0100 |
| commit | ed22bf69917479034aa4d2d42cbd9cb9d59cc0ae (patch) | |
| tree | fbf0fc581d33969e13a77db6383def4c66feff91 /llvm/lib/CodeGen/TargetLoweringObjectFileImpl.cpp | |
| parent | 027935d3cd2886d3dd07741539305f972c5f7e5f (diff) | |
| download | llvm-ed22bf69917479034aa4d2d42cbd9cb9d59cc0ae.zip llvm-ed22bf69917479034aa4d2d42cbd9cb9d59cc0ae.tar.gz llvm-ed22bf69917479034aa4d2d42cbd9cb9d59cc0ae.tar.bz2 | |
[mlir][linalg] Fix weight dimension ordering in 2D grouped conv (#73855)
The `conv_2d_ngchw_fgchw` Op implements 2d grouped convolution with
dimensions ordered as given in the name. However, the current
implementation orders weights as `gfchw` instead of `fgchw`. This was
already pointed out in an old phabricator revision which never landed:
https://reviews.llvm.org/D150064
This patch
1) Adds a new op `conv_2d_ngchw_gfchw`
2) Fixes the dimension ordering of the old op `conv_2d_ngchw_fgchw`
3) Adds tests with non-dynamic dimensions so that it's easier to
understand.
Diffstat (limited to 'llvm/lib/CodeGen/TargetLoweringObjectFileImpl.cpp')
0 files changed, 0 insertions, 0 deletions