PyTorch 是一個基於 Python 的科學計算包,主要定位兩類人羣:
python
from future import print_function
import torch
x = torch.empty(5, 3)
print(x)
輸出:linux
tensor(1.00000e-04 *
[[-0.0000, 0.0000, 1.5135],
[ 0.0000, 0.0000, 0.0000],
[ 0.0000, 0.0000, 0.0000],
[ 0.0000, 0.0000, 0.0000],
[ 0.0000, 0.0000, 0.0000]])
構造一個隨機初始化的矩陣:
windows
x = torch.rand(5, 3)
print(x)
輸出:ide
tensor([[ 0.6291, 0.2581, 0.6414],
[ 0.9739, 0.8243, 0.2276],
[ 0.4184, 0.1815, 0.5131],
[ 0.5533, 0.5440, 0.0718],
[ 0.2908, 0.1850, 0.5297]])
構造一個矩陣全爲 0,並且數據類型是 long.性能
Construct a matrix filled zeros and of dtype long:
學習
x = torch.zeros(5, 3, dtype=torch.long)
print(x)
輸出:spa
tensor([[ 0, 0, 0],
[ 0, 0, 0],
[ 0, 0, 0],
[ 0, 0, 0],
[ 0, 0, 0]])
構造一個張量,直接使用數據:
code
x = torch.tensor([5.5, 3])
print(x)
輸出:教程
tensor([ 5.5000, 3.0000])
建立一個 tensor 基於已經存在的 tensor。
索引
x = x.new_ones(5, 3, dtype=torch.double)
# new_* methods take in sizes
print(x)
x = torch.randn_like(x, dtype=torch.float)
# override dtype!
print(x)
# result has the same size
輸出:
tensor([[ 1., 1., 1.],
[ 1., 1., 1.],
[ 1., 1., 1.],
[ 1., 1., 1.],
[ 1., 1., 1.]], dtype=torch.float64)
tensor([[-0.2183, 0.4477, -0.4053],
[ 1.7353, -0.0048, 1.2177],
[-1.1111, 1.0878, 0.9722],
[-0.7771, -0.2174, 0.0412],
[-2.1750, 1.3609, -0.3322]])
獲取它的維度信息:
print(x.size())
輸出:
torch.Size([5, 3])
注意
torch.Size
是一個元組,因此它支持左右的元組操做。
y = torch.rand(5, 3)
print(x + y)
Out:
tensor([[-0.1859, 1.3970, 0.5236],
[ 2.3854, 0.0707, 2.1970],
[-0.3587, 1.2359, 1.8951],
[-0.1189, -0.1376, 0.4647],
[-1.8968, 2.0164, 0.1092]])
print(torch.add(x, y))
Out:
tensor([[-0.1859, 1.3970, 0.5236],
[ 2.3854, 0.0707, 2.1970],
[-0.3587, 1.2359, 1.8951],
[-0.1189, -0.1376, 0.4647],
[-1.8968, 2.0164, 0.1092]])
result = torch.empty(5, 3)
torch.add(x, y, out=result)
print(result)
Out:
tensor([[-0.1859, 1.3970, 0.5236],
[ 2.3854, 0.0707, 2.1970],
[-0.3587, 1.2359, 1.8951],
[-0.1189, -0.1376, 0.4647],
[-1.8968, 2.0164, 0.1092]])
# adds x to y
y.add_(x)
print(y)
Out:
tensor([[-0.1859, 1.3970, 0.5236],
[ 2.3854, 0.0707, 2.1970],
[-0.3587, 1.2359, 1.8951],
[-0.1189, -0.1376, 0.4647],
[-1.8968, 2.0164, 0.1092]])
Note
注意 任何使張量會發生變化的操做都有一個前綴 '_'。例如:x.copy_(y)
,
x.t_()
, 將會改變
x
.
你能夠使用標準的 NumPy 相似的索引操做
print(x[:, 1])
Out:
tensor([ 0.4477, -0.0048, 1.0878, -0.2174, 1.3609])
torch.view
:
x = torch.randn(4, 4)
y = x.view(16)
z = x.view(-1, 8) # the size -1 is inferred from other dimensions
print(x.size(), y.size(), z.size())
Out:
torch.Size([4, 4]) torch.Size([16]) torch.Size([2, 8])
x = torch.randn(1)
print(x)
print(x.item())
Out:
tensor([ 0.9422])
0.9422121644020081