torch.Tensor文档学习笔记

A torch.Tensor is a multi-dimensional matrix containing elements of a single data type.

张量(torch.Tensor)是包含单个数据类型元素的多维矩阵。

1、张量定义了如下八种CPU张量类型和八种GPU张量类型：

#CPU对应八种数据类型，GPU对应也有八种数据类型,如torch.cuda.FloatTensor([])
torch.FloatTensor([])
torch.DoubleTensor([])
torch.HalfTensor([])
torch.IntTensor([])
torch.LongTensor([])
torch.ShortTensor([])
torch.ByteTensor([])
torch.CharTensor([])

Data type                    CPU tensor                GPU tensor
32-bit floating point    torch.FloatTensor         torch.cuda.FloatTensor
64-bit floating point    torch.DoubleTensor   torch.cuda.DoubleTensor
16-bit floating point    torch.HalfTensor        torch.cuda.HalfTensor
8-bit integer (unsigned)    torch.ByteTensor        torch.cuda.ByteTensor
8-bit integer (signed)    torch.CharTensor        torch.cuda.CharTensor
16-bit integer (signed)    torch.ShortTensor    torch.cuda.ShortTensor
32-bit integer (signed)    torch.IntTensor         torch.cuda.IntTensor
64-bit integer (signed)    torch.LongTensor         torch.cuda.LongTensor

2、数据的原地修改还是生成副本

#函数后面跟短下划线，表示就地修改
a=torch.Tensor([-1,-2])
b=torch.Tensor([-2,2])
print('a.abs() :',a.abs(),'\n','a:',a)
print('b.abs_() :',b.abs_(),'\n','b:',b)
# a.abs():
# 1
# 2
# [torch.FloatTensor of size 2]
#
# a:
# -1
# -2
# [torch.FloatTensor of size 2]
#
# b.abs_():
# 2
# 2
# [torch.FloatTensor of size 2]
#
# b:
# 2
# 2
# [torch.FloatTensor of size 2]

 

3、backward

 x=torch.tensor([[1,1],[1.,1.]],requires_grad=True)
out=x.pow(2).sum()
out.backward()
x.grad
Out[198]:
tensor([[ 2.,  2.],
        [ 2.,  2.]])