[Mlir-commits] [mlir] [mlir][tosa] Add tosa.argmin lowering to Linalg (PR #233754)

[Jack Frankland]

================
@@ -2691,6 +2796,53 @@ func.func @argmax_nan_ignore(%arg0: tensor<5x4xf32>, %arg1: tensor<5x4xf32>) ->
 
 // -----
 
+// CHECK-LABEL: @argmin_nan_propagate
+func.func @argmin_nan_propagate(%arg0: tensor<5x4xf32>, %arg1: tensor<5x4xf32>) -> () {
+  // CHECK: linalg.generic
----------------
FranklandJack wrote:

Likewise here, should we be capturing these values and checking they are used correctly?

https://github.com/llvm/llvm-project/pull/233754