arr[:,0] arr[...,0]

1. arr[:, 0]相当于arr[:,0,:]

表示取arr第0维全取，第一维只取0，后面的维度全取

arr = torch.tensor([
    [[1, 2, 3, 4], [5, 6, 7, 8],     [9, 10, 11, 12]],
    [[13, 14, 15, 16], [17, 18, 19, 20], [21, 22, 23, 24]]
])
arr1=arr[:, 0]
print(arr1)

 egg:

x=torch.randn(2,3,4)
print(x)
x1=x[:,0, :]
print(x1.shape)
print(x1)

 

2. arr[...,0] == arr[:,:,0]

表示前面的维度全取，后面最后一个维度只取0

egg:

x2=x[...,0]
x3=x[:,:,0]
print(x2.shape)
print(x2)
print(x3.shape)
print(x3)