std :: ml::gpu_rocm
Function std::rocm_half_gemm_relu
int rocm_half_gemm_relu(
struct RocmHalfBuf a,
int ta,
struct RocmHalfBuf b,
int tb,
struct RocmHalfBuf c,
unsigned long m,
unsigned long n,
unsigned long k,
)c = relu(op(a) op(b)), all in a's dtype (the GEMM, then ReLU in place).