amdex
2019-07-03 22:06
如果要使用numpy，则可以使用索引数组，其中包含相对于所需索引的邻居索引，然后将其添加到所需索引。我个人认为这很优雅，但是YMMV
这是一个例子：
import numpy as np
# A 5 * 5 grid
grid = np.arange(25).reshape(5, 5)
# A schematic representation of the grid
# 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
# We define how our neighbors relate to our index.
mask = np.array([[0, 1], [1, 0], [0, -1], [-1, 0]])
# Let's say we want the neighbors of [2, 2], which are 17, 11, 7, and 13
# Index marked with X, neighbors marked with O
# 0, 1, 2, 3, 4
# 5, 6, O 8, 9
# 10, O X O 14
# 15, 16, O 18, 19
# 20, 21, 22, 23, 24
desired_index = np.array([2, 2])
# We add the neighbor indices to the mask
neighbor_indices = desired_index + mask
# [[2, 3], [3, 2], [2, 1], [1, 2]]
# Index the array using the indices.
neighbors = grid[neighbor_indices[:, 0], neighbor_indices[:, 1]]
请注意，此示例并未解决超出范围的问题。具体来说，当给定的索引高于列或行的数目时，它将出错，并且对于小于0的索引将回绕。
desired_index = np.array([0, 0])
neighbor_indices = desired_index + mask
neighbors = grid[neighbor_indices[:, 0], neighbor_indices[:, 1]]
# Wrong
desired_index = np.array([4, 4])
neighbor_indices = desired_index + mask
neighbors = grid[neighbor_indices[:, 0], neighbor_indices[:, 1]]
# Error
热门帖子
相关问题
热门github
1
2
3
5
6
7
8
11
13
14
15
Mamba is a new state space model architecture showing promising performance on information-dense data such as language modeling, where previous subquadratic models fall short of Transformers. It is based on the line of progress on structured state space models, with an efficient hardware-aware design and implementation in the spirit of FlashAttention.
(翻译：Mamba 是一种新的状态空间模型架构，在信息密集型数据（例如语言建模）上显示出良好的性能，而之前的二次模型在 Transformers 方面存在不足。它基于结构化状态空间模型的进展，并本着FlashAttention的精神进行高效的硬件感知设计和实现。)
棒极了。确实非常优雅。尽管我只处理小型的普通2D列表，所以使用numpy有点过分。我的问题确实更多是关于清楚表达我的意图。也许我只是想得太多。
如果您认为数组过大，则仍可以使用类似的方法使用列表。
[grid[desired_x + x, desired_y + y] for x, y in [[-1, 0], [0, -1], [1, 0], [0, 1]]]看起来有些笨拙，但我认为您明白了。无论如何，我认为最初的想法很明确。