/
trainer.py
49 lines (29 loc) · 1.41 KB
/
trainer.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
import math
import tensorflow as tf
import numpy as np
class Trainer():
def __init__(self, sess, model, data_train, batch_size=100):
self.sess = sess
self.model = model
self.data_train = data_train
self.batch_size = batch_size
def train(self, n_epochs, p_epochs=10):
for epoch in range(n_epochs):
train_loss, train_acc = self.train_epoch(epoch)
if (epoch + 1) % p_epochs == 0:
print('Epoch : {:<3d} | Loss : {:.5f} | Train Accuracy : {:.5f}'.format(epoch + 1, train_loss, train_acc))
self.model.save(self.sess)
def train_epoch(self, epoch):
avg_loss = 0
avg_acc = 0
n_itrs = math.ceil(len(self.data_train[0]) / self.batch_size)
for itr in range(n_itrs):
loss, acc = self.train_step(itr)
avg_loss += loss / n_itrs
avg_acc += acc / n_itrs
return avg_loss, avg_acc
def train_step(self, itr):
batch_xs, batch_ys = self.data_train[0][itr * self.batch_size:(itr + 1) * self.batch_size], self.data_train[1][itr * self.batch_size:(itr + 1) * self.batch_size]
feed_dict = {self.model.X: batch_xs, self.model.Y: batch_ys}
_, loss, acc = self.sess.run([self.model.train, self.model.loss, self.model.accuracy], feed_dict=feed_dict)
return loss, acc