[all-commits] [llvm/llvm-project] 4a3d20: [mlir][linalg] Add TransposeConv2D Transform Op (#...

[Jack Frankland via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 4a3d2088d61e994dd1aa4e66cdecd15f28d2f397
      https://github.com/llvm/llvm-project/commit/4a3d2088d61e994dd1aa4e66cdecd15f28d2f397
  Author: Jack Frankland <jack.frankland at arm.com>
  Date:   2023-11-28 (Tue, 28 Nov 2023)

  Changed paths:
    M mlir/include/mlir/Dialect/Linalg/TransformOps/LinalgTransformOps.td
    M mlir/include/mlir/Dialect/Linalg/Transforms/Transforms.h
    M mlir/lib/Dialect/Linalg/TransformOps/LinalgTransformOps.cpp
    M mlir/lib/Dialect/Linalg/Transforms/CMakeLists.txt
    A mlir/lib/Dialect/Linalg/Transforms/TransposeConv2D.cpp
    A mlir/test/Dialect/Linalg/transpose-conv2d.mlir

  Log Message:
  -----------
  [mlir][linalg] Add TransposeConv2D Transform Op (#68567)

* Add a LinAlg pass to convert 2D convolutions and quantized 2D
convolutions that have the `FHWC` filter channel ordering into a
transpose followed by 2D convolutions that have the `HWCF` channel
ordering.

* Add a lit test to check the semantics of the transformation are
correct for both quantized and unquantized variants.

Signed-off-by: Jack Frankland <jack.frankland at arm.com>