## print model summary
model.summary()
# Load model weights
n_trained_chunks = 0
if args.load_state != "":
model.load_weights(args.load_state)
n_trained_chunks = int(re.match(".*epoch([0-9]+).*", args.load_state).group(1))
# Build data generators
train_data_gen = utils.BatchGen(reader=train_reader,
discretizer=discretizer,
normalizer=normalizer,
ihm_pos=args_dict['ihm_pos'],
partition=args.partition,
target_repl=target_repl,
batch_size=args.batch_size,
small_part=args.small_part,
shuffle=True)
val_data_gen = utils.BatchGen(reader=val_reader,
discretizer=discretizer,
normalizer=normalizer,
ihm_pos=args_dict['ihm_pos'],
partition=args.partition,
target_repl=target_repl,
batch_size=args.batch_size,
small_part=args.small_part,
shuffle=False)
if args.mode == 'train':
normalizer_state)
normalizer.load_params(normalizer_state)
tf.logging.info(str(vars(conf)))
tf.logging.info(str(args))
number_epoch = int(args['number_epoch'])
batch_size = int(args['batch_size'])
if args['mode'] in ['train', 'eval']:
sp = True if args['mode'] == 'eval' else conf.small_part
train_data_gen = mt_utils.BatchGen(reader=train_reader,
discretizer=discretizer,
normalizer=normalizer,
batch_size=batch_size,
shuffle=True,
return_names=True,
ihm_pos=48,
partition='custom',
target_repl=False,
small_part=sp)
eval_data_gen = mt_utils.BatchGen(reader=val_reader,
discretizer=discretizer,
normalizer=normalizer,
batch_size=batch_size,
shuffle=True,
return_names=True,
ihm_pos=48,
partition='custom',
target_repl=False,
small_part=conf.small_part)