sliding_window_view 的逆的数字方法

Question

我有原始数组

test

：

from numpy.lib.stride_tricks import sliding_window_view

test = np.array([1,2,3,4,5,6,7,8,9,10,11,12,13,14]).reshape(-1,7) # (batch_size, seq_len) -> (2,7)
slided = sliding_window_view(test, window_shape=(3,), axis=-1)
print(test, test.shape)
print(slided, slided.shape)

输出：

[[ 1  2  3  4  5  6  7]
 [ 8  9 10 11 12 13 14]] (2, 7)

[[[ 1  2  3]
  [ 2  3  4]
  [ 3  4  5]
  [ 4  5  6]
  [ 5  6  7]]

 [[ 8  9 10]
  [ 9 10 11]
  [10 11 12]
  [11 12 13]
  [12 13 14]]] (2, 5, 3)

给定现有的

slided

数组是通过

sliding_window_view

返回形状

(batch_size, num_win, win_len)

计算的，如何重建回原始数组

test

且形状为

(batch_size, seq_len)

？

Answer 1

已知

batch_size

等于

2

:

batch_size=2
inversed = np.unique(slided.reshape(batch_size, -1)).reshape(batch_size, -1)

输出：

[[ 1  2  3  4  5  6  7]
 [ 8  9 10 11 12 13 14]] (2, 7)