1.GRU的原理
![在这里插入图片描述](https://img-blog.csdnimg.cn/20201015093606476.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NjY0OTA1Mg==,size_16,color_FFFFFF,t_70#pic_center)
2.GRU实战
1.layer
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
import tensorflow as tf
import numpy as np
from tensorflow import keras
from tensorflow.keras import layers
tf.random.set_seed(22)
np.random.seed(22)
assert tf.__version__.startswith('2.')
batchsz = 128
total_words = 10000
max_review_len = 80
embedding_len = 100
(x_train, y_train), (x_test, y_test) = keras.datasets.imdb.load_data(num_words=total_words)
x_train = keras.preprocessing.sequence.pad_sequences(x_train, maxlen=max_review_len)
x_test = keras.preprocessing.sequence.pad_sequences(x_test, maxlen=max_review_len)
db_train = tf.data.Dataset.from_tensor_slices((x_train, y_train))
db_train = db_train.shuffle(1000).batch(batchsz, drop_remainder=True)
db_test = tf.data.Dataset.from_tensor_slices((x_test, y_test))
db_test = db_test.batch(batchsz, drop_remainder=True)
print('x_train shape:', x_train.shape, tf.reduce_max(y_train), tf.reduce_min(y_train))
print('x_test shape:', x_test.shape)
class MyRNN