sess.run(init)
p = 0
count = 0
temp_acc = 0
temp_loss = 0
acc_list = []
loss_list = []
epoch_list = []
val_acc = []
val_loss = []
val_epoch_list = []
file_val = open('val.txt', 'a')
# print(len(val_file_list))
print("即将开始读取数据")
data, label, _ = TFRecordReader(train_file, 0)
label = tf.cast(label, tf.int32)
label = tf.one_hot(label, 2)
label = sess.run(label)
# print("hsuchausihcashdckuashckjasgcjsakhdsagfhdsugfoidsbfvjhsdhfsd")
# print(label)
val_data, val_label, _ = TFRecordReader(val_file, 0)
val_data = val_data[:20]
val_label = val_label[:20]
val_label = tf.cast(val_label, tf.int32)
val_label = tf.one_hot(val_label, 2)
val_label = sess.run(val_label)
val_label = val_label.reshape(len(val_label) * 1024, 2)
val_data = val_data.reshape(len(val_data) * 1024, 1, num_input)
# label
optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate)
train_process = optimizer.minimize(loss_op)
# 定义准确率
acc = tf.reduce_mean(
tf.cast(tf.equal(tf.argmax(prediction, 1), tf.argmax(Y, 1)), tf.float32))
# 保存模型
meraged = tf.summary.merge_all()
tf.add_to_collection('loss', loss_op)
tf.add_to_collection('accuracy', acc)
tf.add_to_collection('prediction', prediction)
# 初始化变量
init = tf.global_variables_initializer()
# tf.get_default_graph().finalize()
data, label, _ = TFRecordReader(train_file, 0)
label = tf.cast(label, tf.int32)
label = tf.one_hot(label, 2)
data = tf.reshape(data, [5000, 1024, 30])
val_data, val_label, _ = TFRecordReader(val_file, 0)
val_data = val_data[:20]
val_label = val_label[:20]
val_label = tf.cast(val_label, tf.int32)
val_label = tf.one_hot(val_label, 2)
val_label = tf.reshape(val_label, [20 * 1024, 2])
val_data = tf.reshape(val_data, [20, 1024, num_input])
with tf.Session() as sess:
sess.run(init)
# 读取数据，调整shape，保证输入无误