pytorch pack_padded_sequence 实例

import torch
import numpy as np

input = torch.from_numpy(np.array([[1,2,3,4],[5,6,7,8],[9,10,11,12]]))
length = [4,4,4] # lengths array has to be sorted in decreasing order
result = torch.nn.utils.rnn.pack_padded_sequence(input,lengths=length,batch_first=True)
print(result)

input = torch.randn(8,10,300)
length = [10,10,10,10,10,10,10,10]
perm = torch.LongTensor(range(8))
result = torch.nn.utils.rnn.pack_padded_sequence(input[perm],lengths=length,batch_first=True)
print(result)

PackedSequence(data=
1
5
9
2
6
10
3
7
11
4
8
12
[torch.LongTensor of size 12]
, batch_sizes=[3, 3, 3, 3])

PackedSequence(data=
-1.0129e+00 -1.5844e+00 -4.0759e-02  ...  -9.6837e-01  5.7004e-01 -1.6919e-01
5.4662e-01  8.6405e-01  7.8474e-01  ...   5.2483e-01  1.9581e-02  7.2974e-01
7.7569e-02 -7.1858e-03 -2.9401e-01  ...  -2.5550e-01  6.6782e-01  5.6192e-01
...                   ⋱                   ...
-6.7423e-01 -1.6357e+00  1.4011e+00  ...   6.4557e-02  9.4204e-01  6.1430e-01
-1.0300e+00  4.6429e-01  1.4219e+00  ...   2.9208e+00  1.5081e+00  7.6805e-02
-7.2723e-01  6.2770e-01 -6.2025e-01  ...  -3.5286e-01  1.0199e+00  8.8412e-01
[torch.FloatTensor of size 80x300]
, batch_sizes=[8, 8, 8, 8, 8, 8, 8, 8, 8, 8])