pytorch数据类型介绍

用法介绍

 pytorch的数据类型比较繁杂，有tensor内的数据类型，tensor在CPU中的数据类型以及tensor在GPU中的数据类型，它们之间的对应关系如下表格所示

Data type	type	CPU tensor	GPU tensor
64-bit floating point	torch.float64 /torch.double	torch.DoubleTensor	torch.cuda.DoubleTensor
32-bit floating point	torch.float16 /torch.half	torch.HalfTensor	torch.cuda.HalfTensor
8-bit integer (unsigned)	torch.uint8	torch.ByteTensor	torch.cuda.ByteTensor
8-bit integer (signed)	torch.int8	torch.CharTensor	torch.cuda.CharTensor
16-bit integer (signed)	torch.int16/torch.short	torch.ShortTensor	torch.cuda.ShortTensor
32-bit integer (signed)	torch.int32/torch.int	torch.IntTensor	torch.cuda.IntTensor
64-bit integer (signed)	torch.int64/torch.long	torch.LongTensor	torch.cuda.LongTensor
Boolean	torch.bool	torch.BoolTensor	torch.cuda.BoolTensor

 pytorch中对应的数据类型转换函数有tensor1.type_as(tensor2)，tensor.type(torch.IntTensor)，tensor.long()，tensor.char()，tensor.int()，tensor.byte()，tensor.double()，tenosr.to(torch.long)

代码实例

 pytorch的数据类型定义以及对应的数据类型转换的程序如下所示

>>> import torch
>>> torch.zeros((2, 4), dtype=torch.int32)
tensor([[0, 0, 0, 0],
        [0, 0, 0, 0]], dtype=torch.int32)
>>> torch.zeros((2, 4), dtype=torch.float32)
tensor([[0., 0., 0., 0.],
        [0., 0., 0., 0.]])
>>> torch.zeros((2, 4), dtype=torch.uint8)
tensor([[0, 0, 0, 0],
        [0, 0, 0, 0]], dtype=torch.uint8)
>>> torch.zeros((2, 4), dtype=torch.float64)
tensor([[0., 0., 0., 0.],
        [0., 0., 0., 0.]], dtype=torch.float64)
>>> A = torch.zeros((2, 4), dtype=torch.int32)
>>> type(A)
<class 'torch.Tensor'>
>>> B = torch.zeros((2, 4), dtype=torch.float32).type_as(A)
>>> print(B)
tensor([[0, 0, 0, 0],
        [0, 0, 0, 0]], dtype=torch.int32)