在 numpy/pytorch 中取多个切片
Take multiply slices in numpy/pytorch
我有一个大的一维数组 X.shape = (10000,) 和一个索引向量 y = [0, 7, 9995]。
我想得到一个包含行的矩阵
[
X[0 : 100],
X[7 : 107],
concat(X[9995:], X[:95]),
]
也就是说，长度为 100 的切片，从每个索引开始，环绕。
我可以用 python 循环来做到这一点，但我想知道在 pytorch 或 numpy 中是否有更智能的批处理方法，因为我的数组可能非常大。
其实很简单。
- 对于
y 中的每个元素 E，创建一个从 E 到 E + 100 的范围
- 水平连接所有范围
- 用
X 的长度对结果数组求模
indexes = np.hstack([np.arange(v, v + 100) for v in y]) % X.shape[0]
输出：
>>> indexes
array([ 0, 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, 29, 30, 31, 32,
33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43,
44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54,
55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65,
66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76,
77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87,
88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98,
99, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27,
28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38,
39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49,
50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60,
61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71,
72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82,
83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93,
94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104,
105, 106, 9995, 9996, 9997, 9998, 9999, 0, 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, 29, 30, 31, 32, 33, 34, 35, 36,
37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47,
48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58,
59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69,
70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80,
81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91,
92, 93, 94])
现在只需使用索引 X 即可：
X[indexes]
我有一个大的一维数组 X.shape = (10000,) 和一个索引向量 y = [0, 7, 9995]。
我想得到一个包含行的矩阵
[
X[0 : 100],
X[7 : 107],
concat(X[9995:], X[:95]),
]
也就是说，长度为 100 的切片，从每个索引开始，环绕。
我可以用 python 循环来做到这一点，但我想知道在 pytorch 或 numpy 中是否有更智能的批处理方法，因为我的数组可能非常大。
其实很简单。
- 对于
y中的每个元素E，创建一个从E到E + 100 的范围
- 水平连接所有范围
- 用
X 的长度对结果数组求模
indexes = np.hstack([np.arange(v, v + 100) for v in y]) % X.shape[0]
输出：
>>> indexes
array([ 0, 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, 29, 30, 31, 32,
33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43,
44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54,
55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65,
66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76,
77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87,
88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98,
99, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27,
28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38,
39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49,
50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60,
61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71,
72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82,
83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93,
94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104,
105, 106, 9995, 9996, 9997, 9998, 9999, 0, 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, 29, 30, 31, 32, 33, 34, 35, 36,
37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47,
48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58,
59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69,
70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80,
81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91,
92, 93, 94])
现在只需使用索引 X 即可：
X[indexes]