sess = tf.Session(config=session_conf)
with sess.as_default():
checkpoint_file = FLAGS.checkpoint_file
saver = tf.train.import_meta_graph("{}.meta".format(checkpoint_file))
saver.restore(sess, checkpoint_file)
input_x = graph.get_operation_by_name('input_x').outputs[0]
dropout_keep_prob = graph.get_operation_by_name(
'dropout_keep_prob').outputs[0]
predictions = graph.get_operation_by_name('predictions').outputs[0]
batches = batch_iter(list(x_test), FLAGS.batch_size, shuffle=False)
for x_batch in batches:
cand_predictions = sess.run(predictions, {
input_x: x_batch,
dropout_keep_prob: 1.0
})
all_predictions = np.concatenate(
(all_predictions, cand_predictions))
print y_test[0]
print all_predictions[0]
print type(all_predictions)
def dev_step(x_batch, y_batch):
feed_dict = {
rnn.input_x: x_batch,
rnn.input_y: y_batch,
rnn.dropout_keep_prob: 1.0
}
step, loss, accuracy = sess.run(
[rnn.global_step, rnn.loss_val, rnn.accuracy], feed_dict)
time_str = datetime.datetime.now().isoformat()
print "dev_result: {}:step {}, loss {:g}, acc {:g}".format(
time_str, step, loss, accuracy)
for epoch_idx in range(FLAGS.num_epochs):
batches = batch_iter(list(zip(x_train, y_train)), FLAGS.batch_size)
for batch in batches:
x_batch, y_batch = zip(*batch)
train_step(x_batch, y_batch)
if epoch_idx % FLAGS.validate_every == 0:
print '\n'
dev_step(x_dev, y_dev)
path = saver.save(sess,
checkpoints_prefix,
global_step=epoch_idx)
print("Saved model checkpoint to {}\n".format(path))