torch:mul
(torch:mul a b)
Differentiable elementwise (Hadamard) a * b with numpy-style broadcasting (linalg:mul); the matrix product is torch:matmul. Each operand's gradient is the incoming gradient times the other operand.
(torch:mul a b)
Differentiable elementwise (Hadamard) a * b with numpy-style broadcasting (linalg:mul); the matrix product is torch:matmul. Each operand's gradient is the incoming gradient times the other operand.