[Mlir-commits] [mlir] [mlir] Fix semantics of linalg::ReduceOp with several inputs (PR #107005)

[Clément Fournier]

oowekyala wrote:

I wasn't aware of this restructuring effort of linalg. It seems the new `contract` operation would fit my use case most of the time (although it is binary and not variadic)...

I will restrict the verifier in this PR to make `reduce` only unary.

https://github.com/llvm/llvm-project/pull/107005