机器学习 - Indexing 使用

有时候，你需要从tensor里取出特定的数据，这就得用到 indexing 的方法了。

直接上代码

import torch 

x = torch.arange(1, 13).reshape(1, 4, 3) # 1个3维，4个inner array, 每个inner array里有3个元素
print(x)
print(x.shape)

# 结果如下
tensor([[[ 1,  2,  3],
         [ 4,  5,  6],
         [ 7,  8,  9],
         [10, 11, 12]]])
torch.Size([1, 4, 3])

1
2
3
4
5
6
7
8
9
10
11
12
13

获取index对应的位置

print(f"First square bracket:\n {x[0]}")
print(f"Second square bracket: \n {x[0][0]}")
print(f"Third square bracket: \n  {x[0][0][0]}")

# Get all values of 0th dimension and the 0 index of 1st dimension
print(f"Get 0 index of 1st dimension: {x[:, 0]}")
# Get all values of 0th & 1st dimensions but only index 1 of 2nd dimension
print(f"Get all values of 0th & 1st dimension: {x[:, :, 1]}")
# Get all values of the 0 dimension but only the 1 index value of the 1st and 2nd dimension
print(f"Get all values of the 0 dimension but only the 1 index value: {x[:, 1, 1]}")
# Get index 0 of 0th and 1st dimension and all values of 2nd dimension
print(f"Get index 0 of 0th and 1st dimension: {x[0, 0, :]}") # 跟 x[0][0] 是一样的

# 结果如下
First square bracket:
 tensor([[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9],
        [10, 11, 12]])
Second square bracket: 
 tensor([1, 2, 3])
Third square bracket: 
  1
Get 0 index of 1st dimension: tensor([[1, 2, 3]])
Get all values of 0th & 1st dimension: tensor([[ 2,  5,  8, 11]])
Get all values of the 0 dimension but only the 1 index value: tensor([5])
Get index 0 of 0th and 1st dimension: tensor([1, 2, 3])

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

看到这，给个赞呗~