def main(model_path, epoch):
# Load config file
with open(os.path.join(model_path, 'config.yml'), "r") as f:
config = yaml.load(f)
param = config['param']
if param['label_type'] == 'phone61':
param['num_classes'] = 63
param['sos_index'] = 0
param['eos_index'] = 1
elif param['label_type'] == 'phone48':
param['num_classes'] = 50
param['sos_index'] = 0
param['eos_index'] = 1
elif param['label_type'] == 'phone39':
param['num_classes'] = 41
param['sos_index'] = 0
param['eos_index'] = 1
elif param['label_type'] == 'character':
param['num_classes'] = 33
param['sos_index'] = 1
param['eos_index'] = 2
# Model setting
# AttentionModel = load(model_type=param['model'])
network = blstm_attention_seq2seq.BLSTMAttetion(
batch_size=1,
input_size=param['input_size'],
encoder_num_unit=param['encoder_num_unit'],
encoder_num_layer=param['encoder_num_layer'],
attention_dim=param['attention_dim'],
attention_type=param['attention_type'],
decoder_num_unit=param['decoder_num_unit'],
decoder_num_layer=param['decoder_num_layer'],
embedding_dim=param['embedding_dim'],
num_classes=param['num_classes'],
sos_index=param['sos_index'],
eos_index=param['eos_index'],
max_decode_length=param['max_decode_length'],
attention_smoothing=param['attention_smoothing'],
attention_weights_tempareture=param['attention_weights_tempareture'],
logits_tempareture=param['logits_tempareture'],
parameter_init=param['weight_init'],
clip_grad=param['clip_grad'],
clip_activation_encoder=param['clip_activation_encoder'],
clip_activation_decoder=param['clip_activation_decoder'],
dropout_ratio_input=param['dropout_input'],
dropout_ratio_hidden=param['dropout_hidden'],
weight_decay=param['weight_decay'])
network.model_dir = model_path
print(network.model_dir)
do_decode(network=network, param=param, epoch=epoch)
def main(model_path):
epoch = None # if None, restore the final epoch
# Load config file
with open(os.path.join(model_path, 'config.yml'), "r") as f:
config = yaml.load(f)
corpus = config['corpus']
feature = config['feature']
param = config['param']
if corpus['label_type'] == 'phone61':
output_size = 63
elif corpus['label_type'] == 'phone48':
output_size = 50
elif corpus['label_type'] == 'phone39':
output_size = 41
elif corpus['label_type'] == 'character':
output_size = 33
# Model setting
# AttentionModel = load(model_type=config['model_name'])
network = blstm_attention_seq2seq.BLSTMAttetion(
batch_size=1,
input_size=feature['input_size'],
encoder_num_unit=param['encoder_num_unit'],
encoder_num_layer=param['encoder_num_layer'],
attention_dim=param['attention_dim'],
decoder_num_unit=param['decoder_num_unit'],
decoder_num_layer=param['decoder_num_layer'],
embedding_dim=param['embedding_dim'],
output_size=output_size,
sos_index=output_size - 2,
eos_index=output_size - 1,
max_decode_length=param['max_decode_length'],
attention_weights_tempareture=param['attention_weights_tempareture'],
logits_tempareture=param['logits_tempareture'],
parameter_init=param['weight_init'],
clip_grad=param['clip_grad'],
clip_activation_encoder=param['clip_activation_encoder'],
clip_activation_decoder=param['clip_activation_decoder'],
dropout_ratio_input=param['dropout_input'],
dropout_ratio_hidden=param['dropout_hidden'],
weight_decay=param['weight_decay'])
network.model_dir = model_path
print(network.model_dir)
do_eval(network=network,
label_type=corpus['label_type'],
eos_index=output_size - 1,
epoch=epoch)
def main():
args = parser.parse_args()
# Load config file
with open(os.path.join(args.model_path, 'config.yml'), "r") as f:
config = yaml.load(f)
params = config['param']
params['sos_index'] = 0
params['eos_index'] = 1
if params['label_type'] == 'phone61':
params['num_classes'] = 63
elif params['label_type'] == 'phone48':
params['num_classes'] = 50
elif params['label_type'] == 'phone39':
params['num_classes'] = 41
elif params['label_type'] == 'character':
params['num_classes'] = 30
elif params['label_type'] == 'character_capital_divide':
params['num_classes'] = 74
# Model setting
# AttentionModel = load(model_type=params['model'])
model = blstm_attention_seq2seq.BLSTMAttetion(
input_size=params['input_size'],
encoder_num_unit=params['encoder_num_unit'],
encoder_num_layer=params['encoder_num_layer'],
attention_dim=params['attention_dim'],
attention_type=params['attention_type'],
decoder_num_unit=params['decoder_num_unit'],
decoder_num_layer=params['decoder_num_layer'],
embedding_dim=params['embedding_dim'],
num_classes=params['num_classes'],
sos_index=params['sos_index'],
eos_index=params['eos_index'],
max_decode_length=params['max_decode_length'],
attention_smoothing=params['attention_smoothing'],
attention_weights_tempareture=params['attention_weights_tempareture'],
logits_tempareture=params['logits_tempareture'],
parameter_init=params['weight_init'],
clip_grad=params['clip_grad'],
clip_activation_encoder=params['clip_activation_encoder'],
clip_activation_decoder=params['clip_activation_decoder'],
weight_decay=params['weight_decay'],
beam_width=1)
model.save_path = args.model_path
do_decode(model=model, params=params, epoch=args.epoch)
def main(config_path, model_save_path):
# Load a config file (.yml)
with open(config_path, "r") as f:
config = yaml.load(f)
params = config['param']
params['sos_index'] = 0
params['eos_index'] = 1
if params['label_type'] == 'phone61':
params['num_classes'] = 63
elif params['label_type'] == 'phone48':
params['num_classes'] = 50
elif params['label_type'] == 'phone39':
params['num_classes'] = 41
elif params['label_type'] == 'character':
params['num_classes'] = 30
elif params['label_type'] == 'character_capital_divide':
params['num_classes'] = 74
# Model setting
# AttentionModel = load(model_type=config['model_name'])
model = blstm_attention_seq2seq.BLSTMAttetion(
input_size=params['input_size'],
encoder_num_unit=params['encoder_num_unit'],
encoder_num_layer=params['encoder_num_layer'],
attention_dim=params['attention_dim'],
attention_type=params['attention_type'],
decoder_num_unit=params['decoder_num_unit'],
decoder_num_layer=params['decoder_num_layer'],
embedding_dim=params['embedding_dim'],
num_classes=params['num_classes'],
sos_index=params['sos_index'],
eos_index=params['eos_index'],
max_decode_length=params['max_decode_length'],
attention_smoothing=params['attention_smoothing'],
attention_weights_tempareture=params['attention_weights_tempareture'],
logits_tempareture=params['logits_tempareture'],
parameter_init=params['weight_init'],
clip_grad=params['clip_grad'],
clip_activation_encoder=params['clip_activation_encoder'],
clip_activation_decoder=params['clip_activation_decoder'],
weight_decay=params['weight_decay'],
beam_width=1)
# Set process name
setproctitle('timit_' + model.name + '_' + params['label_type'])
model.name = params['model']
model.name += '_encoder' + str(params['encoder_num_unit'])
model.name += '_' + str(params['encoder_num_layer'])
model.name += '_attdim' + str(params['attention_dim'])
model.name += '_decoder' + str(params['decoder_num_unit'])
model.name += '_' + str(params['decoder_num_layer'])
model.name += '_' + params['optimizer']
model.name += '_lr' + str(params['learning_rate'])
model.name += '_' + params['attention_type']
if bool(params['attention_smoothing']):
model.name += '_smoothing'
if params['attention_weights_tempareture'] != 1:
model.name += '_sharpening' + \
str(params['attention_weights_tempareture'])
if params['weight_decay'] != 0:
model.name += '_weightdecay' + str(params['weight_decay'])
# Set save path
model.save_path = mkdir_join(
model_save_path, 'attention', params['label_type'], model.name)
# Reset model directory
model_index = 0
new_model_path = model.save_path
while True:
if isfile(join(new_model_path, 'complete.txt')):
# Training of the first model have been finished
model_index += 1
new_model_path = model.save_path + '_' + str(model_index)
elif isfile(join(new_model_path, 'config.yml')):
# Training of the first model have not been finished yet
# tf.gfile.DeleteRecursively(new_model_path)
# tf.gfile.MakeDirs(new_model_path)
# break
model_index += 1
new_model_path = model.save_path + '_' + str(model_index)
else:
break
model.save_path = mkdir(new_model_path)
# Save config file
shutil.copyfile(config_path, join(model.save_path, 'config.yml'))
# sys.stdout = open(join(model.save_path, 'train.log'), 'w')
# TODO(hirofumi): change to logger
do_train(model=model, params=params)
def main(config_path):
# Load a config file (.yml)
with open(config_path, "r") as f:
config = yaml.load(f)
corpus = config['corpus']
feature = config['feature']
param = config['param']
if corpus['label_type'] == 'phone61':
output_size = 63
elif corpus['label_type'] == 'phone48':
output_size = 50
elif corpus['label_type'] == 'phone39':
output_size = 41
elif corpus['label_type'] == 'character':
output_size = 33
# Model setting
# AttentionModel = load(model_type=config['model_name'])
network = blstm_attention_seq2seq.BLSTMAttetion(
batch_size=param['batch_size'],
input_size=feature['input_size'],
encoder_num_unit=param['encoder_num_unit'],
encoder_num_layer=param['encoder_num_layer'],
attention_dim=param['attention_dim'],
decoder_num_unit=param['decoder_num_unit'],
decoder_num_layer=param['decoder_num_layer'],
embedding_dim=param['embedding_dim'],
output_size=output_size,
sos_index=output_size - 2,
eos_index=output_size - 1,
max_decode_length=param['max_decode_length'],
attention_weights_tempareture=param['attention_weights_tempareture'],
logits_tempareture=param['logits_tempareture'],
parameter_init=param['weight_init'],
clip_grad=param['clip_grad'],
clip_activation_encoder=param['clip_activation_encoder'],
clip_activation_decoder=param['clip_activation_decoder'],
dropout_ratio_input=param['dropout_input'],
dropout_ratio_hidden=param['dropout_hidden'],
weight_decay=param['weight_decay'])
network.model_name = config['model_name'].upper()
network.model_name += '_encoder' + str(param['encoder_num_unit'])
network.model_name += '_' + str(param['encoder_num_layer'])
network.model_name += '_attdim' + str(param['attention_dim'])
network.model_name += '_decoder' + str(param['decoder_num_unit'])
network.model_name += '_' + str(param['decoder_num_layer'])
network.model_name += '_' + param['optimizer']
network.model_name += '_lr' + str(param['learning_rate'])
if param['weight_decay'] != 0:
network.model_name += '_weightdecay' + str(param['weight_decay'])
# Set save path
network.model_dir = mkdir('/n/sd8/inaguma/result/timit/attention/')
network.model_dir = mkdir_join(network.model_dir, corpus['label_type'])
network.model_dir = mkdir_join(network.model_dir, network.model_name)
# Reset model directory
if not isfile(join(network.model_dir, 'complete.txt')):
tf.gfile.DeleteRecursively(network.model_dir)
tf.gfile.MakeDirs(network.model_dir)
else:
raise ValueError('File exists.')
# Set process name
setproctitle('attention_timit_' + corpus['label_type'])
# Save config file
shutil.copyfile(config_path, join(network.model_dir, 'config.yml'))
sys.stdout = open(join(network.model_dir, 'train.log'), 'w')
print(network.model_name)
do_train(network=network,
optimizer=param['optimizer'],
learning_rate=param['learning_rate'],
batch_size=param['batch_size'],
epoch_num=param['num_epoch'],
label_type=corpus['label_type'],
eos_index=output_size - 1)
sys.stdout = sys.__stdout__
def main(config_path):
# Load a config file (.yml)
with open(config_path, "r") as f:
config = yaml.load(f)
param = config['param']
if param['label_type'] == 'phone61':
param['num_classes'] = 63
param['sos_index'] = 0
param['eos_index'] = 1
elif param['label_type'] == 'phone48':
param['num_classes'] = 50
param['sos_index'] = 0
param['eos_index'] = 1
elif param['label_type'] == 'phone39':
param['num_classes'] = 41
param['sos_index'] = 0
param['eos_index'] = 1
elif param['label_type'] == 'character':
param['num_classes'] = 35
param['sos_index'] = 1
param['eos_index'] = 2
# Model setting
# AttentionModel = load(model_type=config['model_name'])
network = blstm_attention_seq2seq.BLSTMAttetion(
batch_size=param['batch_size'],
input_size=param['input_size'],
encoder_num_unit=param['encoder_num_unit'],
encoder_num_layer=param['encoder_num_layer'],
attention_dim=param['attention_dim'],
attention_type=param['attention_type'],
decoder_num_unit=param['decoder_num_unit'],
decoder_num_layer=param['decoder_num_layer'],
embedding_dim=param['embedding_dim'],
num_classes=param['num_classes'],
sos_index=param['sos_index'],
eos_index=param['eos_index'],
max_decode_length=param['max_decode_length'],
attention_smoothing=param['attention_smoothing'],
attention_weights_tempareture=param['attention_weights_tempareture'],
logits_tempareture=param['logits_tempareture'],
parameter_init=param['weight_init'],
clip_grad=param['clip_grad'],
clip_activation_encoder=param['clip_activation_encoder'],
clip_activation_decoder=param['clip_activation_decoder'],
dropout_ratio_input=param['dropout_input'],
dropout_ratio_hidden=param['dropout_hidden'],
weight_decay=param['weight_decay'])
network.model_name = param['model']
network.model_name += '_encoder' + str(param['encoder_num_unit'])
network.model_name += '_' + str(param['encoder_num_layer'])
network.model_name += '_attdim' + str(param['attention_dim'])
network.model_name += '_decoder' + str(param['decoder_num_unit'])
network.model_name += '_' + str(param['decoder_num_layer'])
network.model_name += '_' + param['optimizer']
network.model_name += '_lr' + str(param['learning_rate'])
network.model_name += '_' + param['attention_type']
if bool(param['attention_smoothing']):
network.model_name += '_smoothing'
if param['attention_weights_tempareture'] != 1:
network.model_name += '_sharpening' + \
str(param['attention_weights_tempareture'])
if param['weight_decay'] != 0:
network.model_name += '_weightdecay' + str(param['weight_decay'])
# Set save path
network.model_dir = mkdir('/n/sd8/inaguma/result/timit/')
network.model_dir = mkdir_join(network.model_dir, 'attention')
network.model_dir = mkdir_join(network.model_dir, param['label_type'])
network.model_dir = mkdir_join(network.model_dir, network.model_name)
# Reset model directory
if not isfile(join(network.model_dir, 'complete.txt')):
tf.gfile.DeleteRecursively(network.model_dir)
tf.gfile.MakeDirs(network.model_dir)
else:
raise ValueError('File exists.')
# Set process name
setproctitle('timit_att_' + param['label_type'])
# Save config file
shutil.copyfile(config_path, join(network.model_dir, 'config.yml'))
sys.stdout = open(join(network.model_dir, 'train.log'), 'w')
print(network.model_name)
do_train(network=network, param=param)