[Mlir-commits] [mlir] [mlir][tosa] Add ERROR_IF checks to TRANSPOSE_CONV2D verifier (PR #133234)

[llvmlistbot at llvm.org]

================
@@ -2896,6 +2896,103 @@ LogicalResult TransposeConv2DOp::inferReturnTypeComponents(
 LogicalResult TransposeConv2DOp::verify() {
   if (verifyConvOp(*this).failed() || verifyConvOpModes(*this).failed())
     return failure();
+
+  const RankedTensorType weightType =
+      llvm::dyn_cast<RankedTensorType>(getWeight().getType());
+  if (!weightType)
+    return success();
+
+  const int64_t kernelHeight = weightType.getDimSize(1);
+  const int64_t kernelWidth = weightType.getDimSize(2);
+
+  // Skip further checks if kernel dimensions are dynamic
+  if (kernelHeight == ShapedType::kDynamic ||
+      kernelWidth == ShapedType::kDynamic)
+    return success();
+
+  llvm::ArrayRef<int64_t> padding = getOutPad();
+  const int64_t outPadTop = padding[0];
----------------
Jerry-Ge wrote:

Other dialects like Linalg (https://mlir.llvm.org/docs/Dialects/Linalg/#linalgconv_2d_nchw_fchw-linalgconv2dnchwfchwop) is using `int64_t` for dilation and stride, etc . Should we change the TOSA dialect? That will be a lot of code change. 

https://github.com/llvm/llvm-project/pull/133234