tensorflow 2以后好多没用的步骤都省略了,嘻嘻。记录一下小bug
1.module ‘tensorflow’ has no attribute ‘assign’
2.module ‘tensorflow’ has no attribute ‘session’
3.module ‘tensorflow’ has no attribute ‘placeholder’
可以直接进行加减乘,无需运行和赋值操作。
3.的解决方案
import tensorflow.compat.v1 as tf
tf.disable_eager_execution()
input1 = tf.placeholder(tf.float32)
4.iris数据集是sklearn中自带的dataset
5.看到有人在网上疑似胡说八道,我就解释一下我理解的卷积就是CBAPD。
C: convolution
B: BN
A: Activation
P: Pooling
D: Dropout