1. Assign a value to a TensorFlow variable
三种方法：
- op = x.assign(new_value)
import tensorflow as tf
x = tf.Variable(12)
x_assign_op = x.assign(34)
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
print("Before assign: ", sess.run(x))
sess.run(x_assign_op)
print("After assign: ", sess.run(x))
"""
output:
Before assign: 12
After assign: 34
"""
- tf.assign(x, new_value)
import tensorflow as tf
x = tf.Variable(12)
y = tf.assign(x, 34)
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
print("Before assign: ", sess.run(x))
sess.run(y)
print("y: ", sess.run(y))
print("After assign: ", sess.run(x))
"""
output:
Before assign: 12
y: 34
After assign: 34
"""
- 直接操作
2. tf.cast
3. argmax
4. tf.one_hot and reshape
5. tf.distributions.Normal
6. Others
tf.squeeze()
squeeze(
input,
axis=None,
name=None,
squeeze_dims=None
)
该函数返回一个张量，这个张量是将原始input中所有维度为1的那些维都删掉的结果。axis可以用来指定要删掉的为1的维度，此处要注意指定的维度必须确保其是1，否则会报错
>>> y = tf.squeeze(inputs, [0, 1], name='squeeze')
>>> ValueError: Can not squeeze dim[0], expected a dimension of 1, got 32 for 'squeeze' (op: 'Squeeze') with input shapes: [32,1,1,3].
例子：
# 't' 是一个维度是[1, 2, 1, 3, 1, 1]的张量
tf.shape(tf.squeeze(t)) # [2, 3]， 默认删除所有为1的维度
# 't' 是一个维度[1, 2, 1, 3, 1, 1]的张量
tf.shape(tf.squeeze(t, [2, 4])) # [1, 2, 3, 1]，标号从零开始，只删掉了2和4维的1