64-bit complex. torch.complex64 or torch.cfloat. 128-bit complex. torch.complex128 or torch.cdouble. 16-bit floating point 1. torch.float16 or torch.half.
確定! 回上一頁