std1.0.0-beta-7 SafeC docs

std :: ml::tensor_half

Function std::tensor_matmul_lowp

#include <std/ml/tensor_half.h> source ยท definition

&FloatTensor tensor_matmul_lowp(const &FloatTensor a, const &FloatTensor b, int dtype)

cast(cast(a) @ cast(b)): operands and result rounded to 'dtype', products accumulated in float32. Gradients flow to a and b in float32.