[Mlir-commits] [mlir] [mlir][tosa] Support more float types in resource transpose folding (PR #213226)

[Sayan Saha]

================
@@ -231,6 +254,19 @@ DenseElementsAttr transpose(ElementsAttr attr, ShapedType inputType,
     if (auto data = tryGetDenseResourceValues<float>(attr);
         data && elementTy.isF32())
       return transposeType(*data, inputType, outputType, permValues);
+
+    if (auto data = tryGetDenseResourceValues<uint8_t>(attr);
+        data && isa<Float4E2M1FNType, Float8E4M3FNType, Float8E5M2Type,
+                    Float8E8M0FNUType>(elementTy))
+      return transposeRawType(*data, inputType, outputType, permValues);
+
+    if (auto data = tryGetDenseResourceValues<uint16_t>(attr);
+        data && isa<Float16Type, BFloat16Type>(elementTy))
+      return transposeRawType(*data, inputType, outputType, permValues);
+
+    if (auto data = tryGetDenseResourceValues<double>(attr);
+        data && elementTy.isF64())
----------------
sahas3 wrote:

It'll be good to add `f64` test point as this is a different code branch.

https://github.com/llvm/llvm-project/pull/213226