def train(net, dataset, epochs, batch_size, print_every=10, show_every=100, figsize=(5,5)):
saver = tf.train.Saver()
sample_z = np.random.uniform(-1, 1, size=(72, z_size))
samples, losses = [], []
steps = 0
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
for e in range(epochs):
for x in dataset.batches(batch_size):
steps += 1
# Sample random noise for G
batch_z = np.random.uniform(-1, 1, size=(batch_size, z_size))
# Run optimizers
_ = sess.run(net.d_opt, feed_dict={net.input_real: x, net.input_z: batch_z})
_ = sess.run(net.g_opt, feed_dict={net.input_z: batch_z, net.input_real: x})
if steps % print_every == 0:
# At the end of each epoch, get the losses and print them out
train_loss_d = net.d_loss.eval({net.input_z: batch_z, net.input_real: x})
train_loss_g = net.g_loss.eval({net.input_z: batch_z})
print("Epoch {}/{}...".format(e+1, epochs),
"Discriminator Loss: {:.4f}...".format(train_loss_d),
"Generator Loss: {:.4f}".format(train_loss_g))
# Save losses to view after training
losses.append((train_loss_d, train_loss_g))
if steps % show_every == 0:
gen_samples = sess.run(
generator(net.input_z, 3, reuse=True, training=False),
feed_dict={net.input_z: sample_z})
samples.append(gen_samples)
_ = view_samples(-1, samples, 6, 12, figsize=figsize)
plt.show()
saver.save(sess, './checkpoints/generator.ckpt')
with open('samples.pkl', 'wb') as f:
pkl.dump(samples, f)
return losses, samples
if steps % show_every == 0:
gen_samples = sess.run(
generator(net.input_z, 3, reuse=True, training=False),
feed_dict={net.input_z: sample_z})
samples.append(gen_samples)
_ = view_samples(-1, samples, 6, 12, figsize=figsize)
plt.show()
saver.save(sess, './checkpoints/generator.ckpt')
with open('samples.pkl', 'wb') as f:
pkl.dump(samples, f)
return losses, samples
# Create the network
net = GAN(real_size, z_size, learning_rate, alpha=alpha, beta1=beta1)
dataset = Dataset(trainset, testset)
losses, samples = train(net, dataset, epochs, batch_size, figsize=(10,5))
_ = view_samples(0, samples, 4, 4, figsize=(10,5))
from os.path import isfile, isdir
from tqdm import tqdm
import os
import cv2
print('Librearies imported %f' % process_time())
from model import GAN
from model import generator
from dataset import view_samples
from constants import *
print('Model and constants imported %f' % process_time())
# Create the network
net = GAN(real_size, z_size, learning_rate, alpha=alpha, beta1=beta1)
print('GAN created %f' % process_time())
saver = tf.train.Saver()
with tf.Session() as sess:
saver.restore(sess, tf.train.latest_checkpoint('checkpoints'))
print('Session restored %f' % process_time())
sample_z = np.random.uniform(-1, 1, size=(int(real_size[0] / 2), z_size))
gen_samples = sess.run(generator(net.input_z,
3,
reuse=True,
training=False),
feed_dict={net.input_z: sample_z})
_ = view_samples(0, [gen_samples], 4, 4, figsize=(5, 5))
def train(net,
dataset,
epochs,
batch_size,
print_every=10,
show_every=100,
figsize=(5, 5)):
print('Training started %f' % process_time())
saver = tf.train.Saver()
sample_z = np.random.uniform(-1, 1, size=(real_size[0] * 2, z_size))
samples, losses = [], []
steps = 0
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
last = process_time()
for e in range(epochs):
for x in dataset.batches(batch_size):
steps += 1
# Sample random noise for G
batch_z = np.random.uniform(-1, 1, size=(batch_size, z_size))
# Run optimizers
_ = sess.run(net.d_opt,
feed_dict={
net.input_real: x,
net.input_z: batch_z
})
_ = sess.run(net.g_opt,
feed_dict={
net.input_z: batch_z,
net.input_real: x
})
if steps % print_every == 0:
# At the end of each epoch, get the losses and print them out
train_loss_d = net.d_loss.eval({
net.input_z: batch_z,
net.input_real: x
})
train_loss_g = net.g_loss.eval({net.input_z: batch_z})
print("Epoch {}/{}...".format(e + 1, epochs),
"Discriminator Loss: {:.4f}...".format(train_loss_d),
"Generator Loss: {:.4f}".format(train_loss_g))
# Save losses to view after training
losses.append((train_loss_d, train_loss_g))
if steps % show_every == 0:
gen_samples = sess.run(generator(net.input_z,
3,
reuse=True,
training=False),
feed_dict={net.input_z: sample_z})
samples.append(gen_samples)
_ = view_samples(-1,
samples,
8,
8,
figsize=figsize,
save=True,
saveCount=len(samples))
saver.save(sess, './checkpoints/generator {}.ckpt'.format(e + 1))
last = process_time() - last
print('Epoch time is %f' % last)
last = process_time()
saver.save(sess, './checkpoints/generator.ckpt')
print('Model trained %f' % process_time())
return losses, samples
return train_accuracies, test_accuracies, samples
real_size = (32, 32, 3)
z_size = 100
learning_rate = 0.0003
net = GAN(real_size, z_size, learning_rate)
dataset = Dataset(trainset, testset)
batch_size = 128
epochs = 25
train_accuracies, test_accuracies, samples = train(net,
dataset,
epochs,
batch_size,
figsize=(10, 5))
fig, ax = plt.subplots()
plt.plot(train_accuracies, label='Train', alpha=0.5)
plt.plot(test_accuracies, label='Test', alpha=0.5)
plt.title("Accuracy")
plt.legend()
_ = view_samples(-1, samples, 5, 10, figsize=(10, 5))
for ii in range(len(samples)):
fig, ax = view_samples(ii, samples, 5, 10, figsize=(10, 5))
fig.savefig('images/samples_{:03d}.png'.format(ii))
plt.close()