def create_model(session, vocab_size, forward_only):
model = nlc_model.NLCModel(
vocab_size, FLAGS.size, FLAGS.num_layers, FLAGS.max_gradient_norm, FLAGS.batch_size,
FLAGS.learning_rate, FLAGS.learning_rate_decay_factor, FLAGS.dropout,
forward_only=forward_only)
ckpt = tf.train.get_checkpoint_state(FLAGS.train_dir)
if ckpt: # and tf.gfile.Exists(ckpt.model_checkpoint_path):
print("Reading model parameters from %s" % ckpt.model_checkpoint_path)
model.saver.restore(session, ckpt.model_checkpoint_path)
else:
print("Created model with fresh parameters.")
session.run(tf.global_variables_initializer())
return model
def create_model(session, vocab_size, forward_only):
model = nlc_model.NLCModel(
vocab_size, FLAGS.size, FLAGS.num_layers, FLAGS.max_gradient_norm, FLAGS.batch_size,
FLAGS.learning_rate, FLAGS.learning_rate_decay_factor, FLAGS.dropout,
forward_only=forward_only, optimizer=FLAGS.optimizer)
ckpt = tf.train.get_checkpoint_state(FLAGS.train_dir)
if ckpt and tf.gfile.Exists(ckpt.model_checkpoint_path):
logging.info("Reading model parameters from %s" % ckpt.model_checkpoint_path)
model.saver.restore(session, ckpt.model_checkpoint_path)
else:
logging.info("Created model with fresh parameters.")
session.run(tf.initialize_all_variables())
logging.info('Num params: %d' % sum(v.get_shape().num_elements() for v in tf.trainable_variables()))
return model
def create_model(session, vocab_size, forward_only):
model = nlc_model.NLCModel(vocab_size,
FLAGS.size,
FLAGS.num_layers,
FLAGS.max_gradient_norm,
FLAGS.batch_size,
FLAGS.learning_rate,
FLAGS.learning_rate_decay_factor,
FLAGS.dropout,
forward_only=forward_only)
ckpt_paths = [f for f in os.listdir(FLAGS.train_dir) if (re.search(r"best\.ckpt-\d+", f) \
and not f.endswith("meta"))]
assert (len(ckpt_paths) > 0)
ckpt_paths = sorted(ckpt_paths, key=lambda x: int(x.split("-")[-1]))
ckpt_path = os.path.join(FLAGS.train_dir, ckpt_paths[-1])
if tf.gfile.Exists(ckpt_path):
print("Reading model parameters from %s" % ckpt_path)
model.saver.restore(session, ckpt_path)
else:
assert (False)
return model
