高阶操作

Outline

• where

• scatter_nd

• meshgrid

Where

where(tensor)

• where获得以下表格中True的位置

1	2	3
True	False	False
False	True	False
False	False	True

import tensorflow as tf

a = tf.random.normal([3, 3])
a

mask = a > 0
mask

# 为True元素的值
tf.boolean_mask(a, mask)

# 为True元素，即>0的元素的索引
indices = tf.where(mask)
indices

# 取回>0的值
tf.gather_nd(a, indices)

where(cond,A,B)

mask

A = tf.ones([3, 3])
B = tf.zeros([3, 3])

# True的元素会从A中选值，False的元素会从B中选值
tf.where(mask, A, B)

scatter_nd

• tf.scatter_nd(
• indices,
• updates,
• shape)

一维

indices = tf.constant([[4], [3], [1], [7]])
updates = tf.constant([9, 10, 11, 12])
shape = tf.constant([8])

# 把updates按照indices的索引放在底板shape上
tf.scatter_nd(indices, updates, shape)

二维

indices = tf.constant([[0], [2]])
updates = tf.constant([
    [[5, 5, 5, 5], [6, 6, 6, 6], [7, 7, 7, 7], [8, 8, 8, 8]],
    [[5, 5, 5, 5], [6, 6, 6, 6], [7, 7, 7, 7], [8, 8, 8, 8]],
])
updates.shape

TensorShape([2, 4, 4])

shape = tf.constant([4, 4, 4])

tf.scatter_nd(indices, updates, shape)

meshgrid

• [-2,-2]
• [-1,-2]
• [0,-2]
• [-2,-2]
• [-1,-1]
• ...
• [2,2]

Points

• [y,x,w]
  • [5,5,2]
• [N,2]

numpy实现

import numpy as np

points = []

for y in np.linspace(-2, 2, 5):
    for x in np.linspace(-2, 2, 5):
        points.append([x, y])

np.array(points)

array([[-2., -2.],
       [-1., -2.],
       [ 0., -2.],
       [ 1., -2.],
       [ 2., -2.],
       [-2., -1.],
       [-1., -1.],
       [ 0., -1.],
       [ 1., -1.],
       [ 2., -1.],
       [-2.,  0.],
       [-1.,  0.],
       [ 0.,  0.],
       [ 1.,  0.],
       [ 2.,  0.],
       [-2.,  1.],
       [-1.,  1.],
       [ 0.,  1.],
       [ 1.,  1.],
       [ 2.,  1.],
       [-2.,  2.],
       [-1.,  2.],
       [ 0.,  2.],
       [ 1.,  2.],
       [ 2.,  2.]])

tensorflow2实现

y = tf.linspace(-2., 2, 5)
y

x = tf.linspace(-2., 2, 5)
x

points_x, points_y = tf.meshgrid(x, y)
points_x.shape

TensorShape([5, 5])

points_x

points_y

points = tf.stack([points_x, points_y], axis=2)
points