def get_create_inputs(dataset_name: str, is_train: bool, epochs: int):
options = {
'mnist': lambda: create_inputs_mnist(is_train),
'fashion_mnist': lambda: create_inputs_mnist(is_train),
'smallNORB': lambda: create_inputs_norb(is_train, epochs)
}
return options[dataset_name]
def get_create_inputs(dataset_name: str, is_train: bool, epochs: int):
options = {'mnist': lambda: create_inputs_mnist(is_train),
'asl': lambda: create_inputs_asl(is_train),
'italy': lambda: create_inputs_italy(is_train),
'fashion_mnist': lambda: create_inputs_mnist(is_train),
'smallNORB': lambda: create_inputs_norb(is_train, epochs),
'cifar10': lambda: create_inputs_cifar10(is_train)}
return options[dataset_name]
def main(_):
coord_add = [[[8., 8.], [12., 8.], [16., 8.]],
[[8., 12.], [12., 12.], [16., 12.]],
[[8., 16.], [12., 16.], [16., 16.]]]
coord_add = np.array(coord_add, dtype=np.float32) / 28.
with tf.Graph().as_default(), tf.device('/cpu:0'):
global_step = tf.get_variable(
'global_step', [], initializer=tf.constant_initializer(0), trainable=False)
num_batches_per_epoch = int(60000 / cfg.batch_size)
opt = tf.train.AdamOptimizer()
batch_x, batch_labels = create_inputs_norb(is_train=True)
# batch_y = tf.one_hot(batch_labels, depth=10, axis=1, dtype=tf.float32)
m_op = tf.placeholder(dtype=tf.float32, shape=())
with tf.device('/gpu:0'):
with slim.arg_scope([slim.variable], device='/cpu:0'):
output = net.build_arch(batch_x, coord_add, is_train=True)
# loss = net.cross_ent_loss(output, batch_labels)
loss = net.spread_loss(output, batch_labels, m_op)
grad = opt.compute_gradients(loss)
loss_name = 'spread_loss'
summaries = []
summaries.append(tf.summary.scalar(loss_name, loss))
train_op = opt.apply_gradients(grad, global_step=global_step)
sess = tf.Session(config=tf.ConfigProto(
allow_soft_placement=True, log_device_placement=False))
sess.run(tf.global_variables_initializer())
saver = tf.train.Saver(tf.global_variables(), max_to_keep=cfg.epoch)
# read snapshot
# latest = os.path.join(cfg.logdir, 'model.ckpt-4680')
# saver.restore(sess, latest)
summary_op = tf.summary.merge(summaries)
tf.train.start_queue_runners(sess=sess)
summary_writer = tf.summary.FileWriter(cfg.logdir, graph=sess.graph)
m = 0.2
for step in range(cfg.epoch * num_batches_per_epoch):
tic = time.time()
_, loss_value = sess.run([train_op, loss], feed_dict={m_op: m})
print('%d iteration is finished in ' % step + '%f second' % (time.time() - tic))
# test1_v = sess.run(test2)
# if np.isnan(loss_value):
# print('bbb')
# assert not np.isnan(np.any(test2_v[0])), 'a is nan'
assert not np.isnan(loss_value), 'loss is nan'
if step % 10 == 0:
summary_str = sess.run(summary_op, feed_dict={m_op: m})
summary_writer.add_summary(summary_str, step)
if (step % num_batches_per_epoch) == 0:
if step > 0:
m += (0.9 - 0.2) / (cfg.epoch * 0.6)
if m > 0.9:
m = 0.9
ckpt_path = os.path.join(cfg.logdir, 'model.ckpt')
saver.save(sess, ckpt_path, global_step=step)
