/
train.py
49 lines (36 loc) · 1.25 KB
/
train.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 tensorflow as tf
tf.enable_eager_execution()
import matplotlib.pyplot as plt
import math, cv2, glob, time
import numpy as np
import imageio as io
from load_data import load_data
from utils import *
import os
patch_size = 256
epochs = 80
bs = 1
ch = 6
def train(model):
dataset = load_data(bs, patch_size)
loss_history = []
print(model.summary())
for epoch in range(epochs):
loss1 = []
step = 0
for ldr, hdr in dataset:
ldr = tf.image.resize_images(tf.cast(ldr, tf.float32), (patch_size, patch_size))*2.0-1.0
hdr = tf.image.resize_images(tf.cast(hdr, tf.float32), (patch_size, patch_size))*2.0-1.0
l = model.train_on_batch(ldr, hdr)
loss1.append(l)
print('epoch:%d, step:%d, model_loss:%f'%(epoch, step, l))
step = step+1
loss_mean = np.mean(loss1)
loss_history.append(loss_mean)
fig = plt.figure()
plt.plot(loss_history)
fig.savefig('loss_history.png', dpi=fig.dpi)
model.save_weights('./lightfuse_model_weights.h5')
if __name__ == "__main__":
model = supervised_model(patch_size, patch_size, ch)
train(model)