Pytorch学习笔记一

本文为根据官方文档做的入门学习笔记

 

• Tensors（张量）

张量就像python中的ndarrays，又可以用于GPU加速计算。

#uninitialized matrix
x = torch.empty(5, 3)
print(x)

tensor([[1.0880e-19, 1.0874e-19, 2.5443e+30],
        [9.4727e+21, 2.4835e+27, 2.5428e+30],
        [1.0877e-19, 1.5163e+23, 2.2012e+12],
        [3.7899e+22, 5.2480e+05, 1.0175e+31],
        [9.7056e+24, 1.6283e+32, 3.7913e+22]])

#randomly initialized matrix
x = torch.rand(5, 3)
print(x)

tensor([[0.1484, 0.8227, 0.9398],
        [0.6957, 0.7653, 0.8424],
        [0.2150, 0.8570, 0.5757],
        [0.1498, 0.6313, 0.4983],
        [0.2018, 0.4304, 0.3919]])

# a matrix filled zeros and of dtype long
x = torch.zeros(5, 3, dtype=torch.long)
print(x)

tensor([[0, 0, 0],
        [0, 0, 0],
        [0, 0, 0],
        [0, 0, 0],
        [0, 0, 0]])

#a tensor directly from data
x = torch.tensor([5.5, 3])
print(x)

tensor([5.5000, 3.0000])

x = x.new_ones(5, 3, dtype=torch.double)
print(x)

x = torch.randn_like(x, dtype=torch.float) #override dtype
print(x)


tensor([[1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.]], dtype=torch.float64)
tensor([[ 0.0412,  0.9730,  1.1397],
        [-0.5710, -1.6817, -0.0902],
        [-0.2467, -1.5928, -0.4624],
        [ 0.3849,  0.2497, -1.3642],
        [ 1.1823,  1.6061, -1.4947]])

#get its size
print(x.size())

torch.Size([5, 3])

Note: torch.size 是个元组，支持所有元组操作

其他语法：

#1
y = torch.rand(5, 3)
print(x + y)

out:
tensor([[ 0.9242,  1.3102,  2.0905],
        [ 0.0236, -1.2898,  0.7076],
        [-0.1685, -1.0089, -0.0096],
        [ 1.1552,  1.0442, -0.4530],
        [ 1.2622,  1.7581, -1.3049]])

#2
print(torch.add(x, y))

out:
tensor([[ 0.9242,  1.3102,  2.0905],
        [ 0.0236, -1.2898,  0.7076],
        [-0.1685, -1.0089, -0.0096],
        [ 1.1552,  1.0442, -0.4530],
        [ 1.2622,  1.7581, -1.3049]])

#3
result = torch.empty(5, 3)
torch.add(x, y, out=result)
print(result)

out:
tensor([[ 0.9242,  1.3102,  2.0905],
        [ 0.0236, -1.2898,  0.7076],
        [-0.1685, -1.0089, -0.0096],
        [ 1.1552,  1.0442, -0.4530],
        [ 1.2622,  1.7581, -1.3049]])

#4
y.add_(x)
print(y)

out:
tensor([[ 0.9242,  1.3102,  2.0905],
        [ 0.0236, -1.2898,  0.7076],
        [-0.1685, -1.0089, -0.0096],
        [ 1.1552,  1.0442, -0.4530],
        [ 1.2622,  1.7581, -1.3049]])

Note: 任何改变一个张量的操作都要加后缀“_”，如x.copy_(y), x.t_(y)都会改变x。

#use torch.view to resize/reshape a tensor
x = torch.randn(4, 4)
y = x.view(16)
z = x.view(-1, 8) #-1 is inferred from other dimension
print(x.size(), y.size(), z.size())

out:
torch.Size([4, 4]) torch.Size([16]) torch.Size([2, 8])

#use .item() to get the value of an one element tensor as a python number
x = torch.randn(1)
print(x)
print(x.item())

out:
tensor([-0.6661])
-0.6661176681518555

tensor和array之间可相互转换(Numpy Bridge)：

tensor和array共享内存位置，改变一个的同时另一个也随着改变。

#convert a tensor to an array
a = torch.ones(5)
print(a)

out:
tensor([1., 1., 1., 1., 1.])

b = a.numpy()
print(b)

out:
[1. 1. 1. 1. 1.]

a.add_(1)
print(a)
print(b)

out:
tensor([2., 2., 2., 2., 2.])
[2. 2. 2. 2. 2.]

#convert an array to a tensor

import numpy as np
a = np.ones(5)
b = torch.from_numpy(a)
np.add(a, 1, out=a)
print(a)
print(b)

out:
[2. 2. 2. 2. 2.]
tensor([2., 2., 2., 2., 2.], dtype=torch.float64)