Matrix mult.

Performs matrix multiplication for the last two dimensions of the inputs.

Takes two inputs, A and B. The supported input data types are Int8, Int16, Int32, Int64, Uint8, Float32 and Float64, with at least two dimensions. The size of the last dimension of A must equal the size of the second last dimension of B. If the shape of A is (*, N, M), and the shape of B is (*, M, P), the output shape will be (*, N, P), where * means any number of dimensions, including zero.