# create a summary to monitor cost tensor
if write_summary:
tf.summary.scalar("loss", loss)
# merge all summaries into a single op
if write_summary:
merged_summary_op = tf.summary.merge_all()
time_start = time.time()
# op to write logs to Tensorboard
if write_summary:
summary_writer = tf.summary.FileWriter(params.save_dir, graph=tf.get_default_graph())
if params.shuffle_training:
data.load_imgs()
# center, curve 50:50%
# data.categorize_imgs()
for i in xrange(params.training_steps):
if params.use_category_normal:
txx, tyy = data.load_batch_category_normal('train')
else:
txx, tyy = data.load_batch('train')
# print "txx: ", len(txx), len(txx[0]), len(txx[0][0]), len(txx[0][0][0])
# print "tyy: ", tyy
train_step.run(feed_dict={model.x: txx, model.y_: tyy})
if write_summary:
tf.summary.scalar("loss", loss)
# merge all summaries into a single op
if write_summary:
merged_summary_op = tf.summary.merge_all()
saver = tf.train.Saver()
time_start = time.time()
# op to write logs to Tensorboard
if write_summary:
summary_writer = tf.summary.FileWriter(params.save_dir, graph=tf.get_default_graph())
if params.shuffle_training:
data.load_imgs()
for i in xrange(params.training_steps):
txx, tyy = data.load_batch('train')
train_step.run(feed_dict={model.x:txx, model.y_:tyy, model.keep_prob: 0.8})
# write logs at every iteration
if write_summary:
summary = merged_summary_op.eval(feed_dict={model.x: txx, model.y_: tyy, model.keep_prob: 1.0})
#summary_writer.add_summary(summary, i)
if (i+1) % 10 == 0:
vxx, vyy = data.load_batch('val')
t_loss = loss.eval(feed_dict={model.x: txx, model.y_: tyy, model.keep_prob: 1.0})
v_loss = loss.eval(feed_dict={model.x: vxx, model.y_: vyy, model.keep_prob: 1.0})