高阶操作

目录

TensorFlow2教程完整教程目录（更有python、go、pytorch、tensorflow、爬虫、人工智能教学等着你）：https://www.cnblogs.com/nickchen121/p/10840284.html

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

<tf.Tensor: id=11, shape=(3, 3), dtype=float32, numpy=
array([[-0.02527909, -0.09084062,  0.34427297],
       [-0.45223615,  1.1085868 , -1.9480664 ],
       [-2.3520288 , -1.8698558 , -0.30862013]], dtype=float32)>

mask = a > 0
mask

<tf.Tensor: id=16, shape=(3, 3), dtype=bool, numpy=
array([[False, False,  True],
       [False,  True, False],
       [False, False, False]])>

# 为True元素的值
tf.boolean_mask(a, mask)

<tf.Tensor: id=44, shape=(2,), dtype=float32, numpy=array([0.34427297, 1.1085868 ], dtype=float32)>

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

<tf.Tensor: id=47, shape=(2, 2), dtype=int64, numpy=
array([[0, 2],
       [1, 1]])>

# 取回>0的值
tf.gather_nd(a, indices)

<tf.Tensor: id=49, shape=(2,), dtype=float32, numpy=array([0.34427297, 1.1085868 ], dtype=float32)>

where(cond,A,B)

mask

<tf.Tensor: id=16, shape=(3, 3), dtype=bool, numpy=
array([[False, False,  True],
       [False,  True, False],
       [False, False, False]])>

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)

<tf.Tensor: id=61, shape=(3, 3), dtype=float32, numpy=
array([[0., 0., 1.],
       [0., 1., 0.],
       [0., 0., 0.]], dtype=float32)>

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)

<tf.Tensor: id=71, shape=(8,), dtype=int32, numpy=array([ 0, 11,  0, 10,  9,  0,  0, 12], dtype=int32)>

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)

<tf.Tensor: id=76, shape=(4, 4, 4), dtype=int32, numpy=
array([[[5, 5, 5, 5],
        [6, 6, 6, 6],
        [7, 7, 7, 7],
        [8, 8, 8, 8]],

       [[0, 0, 0, 0],
        [0, 0, 0, 0],
        [0, 0, 0, 0],
        [0, 0, 0, 0]],

       [[5, 5, 5, 5],
        [6, 6, 6, 6],
        [7, 7, 7, 7],
        [8, 8, 8, 8]],

       [[0, 0, 0, 0],
        [0, 0, 0, 0],
        [0, 0, 0, 0],
        [0, 0, 0, 0]]], dtype=int32)>

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

<tf.Tensor: id=81, shape=(5,), dtype=float32, numpy=array([-2., -1.,  0.,  1.,  2.], dtype=float32)>

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

<tf.Tensor: id=86, shape=(5,), dtype=float32, numpy=array([-2., -1.,  0.,  1.,  2.], dtype=float32)>

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

TensorShape([5, 5])

points_x

<tf.Tensor: id=130, shape=(5, 5), dtype=float32, numpy=
array([[-2., -1.,  0.,  1.,  2.],
       [-2., -1.,  0.,  1.,  2.],
       [-2., -1.,  0.,  1.,  2.],
       [-2., -1.,  0.,  1.,  2.],
       [-2., -1.,  0.,  1.,  2.]], dtype=float32)>

points_y

<tf.Tensor: id=131, shape=(5, 5), dtype=float32, numpy=
array([[-2., -2., -2., -2., -2.],
       [-1., -1., -1., -1., -1.],
       [ 0.,  0.,  0.,  0.,  0.],
       [ 1.,  1.,  1.,  1.,  1.],
       [ 2.,  2.,  2.,  2.,  2.]], dtype=float32)>

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

<tf.Tensor: id=135, shape=(5, 5, 2), dtype=float32, numpy=
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.]]], dtype=float32)>

posted @ 2019-05-14 20:30  B站-水论文的程序猿  阅读(997)  评论(0)  编辑  收藏  举报