One . torch.repeat() Function analysis

1. explain

Official website ：torch.tensor.repeat(), The function description is shown in the following figure ：

2. The functionality

torch.tensor.repeat() Function can repeatedly expand the tensor
1) When there are only two parameters ：（ Repetition multiple of row , Repetition times of columns ）,1 Means not to repeat .
2) When there are three parameters ：（ Repetition multiple of the number of channels , Repetition multiple of row , Repetition times of columns ）,1 Means not to repeat .

3. The code example is as follows ：

3.1 Input one-dimensional tensor , Parameter is one , That is, repeat on the column n Time

a = torch.randn(3)
a,a.repeat(4)

 The results are shown below ：
(tensor([ 0.81, -0.57,  0.10]),
 tensor([ 0.81, -0.57,  0.10,  0.81, -0.57,  0.10,  0.81, -0.57,  0.10,  0.81,
         -0.57,  0.10]))

3.2 Input one-dimensional tensor , The parameters are two (m,n), That is, repeat on the column first n Time , Repeat above the line m Time , The output tensor is two-dimensional

a = torch.randn(3)
a,a.repeat(4,2)

(tensor([ 0.06, -0.76, -0.59]),
 tensor([[ 0.06, -0.76, -0.59,  0.06, -0.76, -0.59],
         [ 0.06, -0.76, -0.59,  0.06, -0.76, -0.59],
         [ 0.06, -0.76, -0.59,  0.06, -0.76, -0.59],
         [ 0.06, -0.76, -0.59,  0.06, -0.76, -0.59]]))

3.3 Input one-dimensional tensor , The parameters are three (b,m,n), That is, repeat on the column first n Time , Repeat above the line m Time , Finally repeat on the channel b Time , The output tensor is three-dimensional

a = torch.randn(3)
a,a.repeat(3,4,2)

 The output is as follows ：
(tensor([2.25, 0.49, 1.47]),
 tensor([[[2.25, 0.49, 1.47, 2.25, 0.49, 1.47],
          [2.25, 0.49, 1.47, 2.25, 0.49, 1.47],
          [2.25, 0.49, 1.47, 2.25, 0.49, 1.47],
          [2.25, 0.49, 1.47, 2.25, 0.49, 1.47]],
 
         [[2.25, 0.49, 1.47, 2.25, 0.49, 1.47],
          [2.25, 0.49, 1.47, 2.25, 0.49, 1.47],
          [2.25, 0.49, 1.47, 2.25, 0.49, 1.47],
          [2.25, 0.49, 1.47, 2.25, 0.49, 1.47]],
 
         [[2.25, 0.49, 1.47, 2.25, 0.49, 1.47],
          [2.25, 0.49, 1.47, 2.25, 0.49, 1.47],
          [2.25, 0.49, 1.47, 2.25, 0.49, 1.47],
          [2.25, 0.49, 1.47, 2.25, 0.49, 1.47]]]))

3.4 Enter the two-dimensional tensor , The parameters are two (m,n), That is, repeat on the column first n Time , Repeat above the line m Time , The output tensor is two-dimensional （ Note that the number of parameters must be greater than the number of input tensor dimensions ）

a = torch.randn(3,2)
a,a.repeat(4,2)

 The output is as follows ：
(tensor([[-0.58, -1.21],
         [-0.35,  0.68],
         [ 0.33,  0.70]]),
 tensor([[-0.58, -1.21, -0.58, -1.21],
         [-0.35,  0.68, -0.35,  0.68],
         [ 0.33,  0.70,  0.33,  0.70],
         [-0.58, -1.21, -0.58, -1.21],
         [-0.35,  0.68, -0.35,  0.68],
         [ 0.33,  0.70,  0.33,  0.70],
         [-0.58, -1.21, -0.58, -1.21],
         [-0.35,  0.68, -0.35,  0.68],
         [ 0.33,  0.70,  0.33,  0.70],
         [-0.58, -1.21, -0.58, -1.21],
         [-0.35,  0.68, -0.35,  0.68],
         [ 0.33,  0.70,  0.33,  0.70]]))

3.5 Enter the two-dimensional tensor , The parameters are three (b,m,n), That is, repeat on the column first n Time , Repeat above the line m Time , Finally repeat on the channel b Time , The output tensor is three-dimensional .（ Note that the number of output tensor dimensions is the number of parameters ）

a = torch.randn(3,2)
a,a.repeat(3,4,2)

 The output is as follows ：
(tensor([[-0.75,  1.20],
         [-1.50,  1.75],
         [-0.05,  0.40]]),
 tensor([[[-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40],
          [-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40],
          [-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40],
          [-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40]],
 
         [[-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40],
          [-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40],
          [-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40],
          [-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40]],
 
         [[-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40],
          [-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40],
          [-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40],
          [-0.75,  1.20, -0.75,  1.20],
          [-1.50,  1.75, -1.50,  1.75],
          [-0.05,  0.40, -0.05,  0.40]]]))