def start(self):
args = self.args
# Set default configuration in args.py
#args = get_args()
logger = self.get_logger()
# Set random seed for reproducibility
torch.manual_seed(args.seed)
torch.backends.cudnn.deterministic = True
np.random.seed(args.seed)
random.seed(args.seed)
if not args.cuda:
args.gpu = -1
if torch.cuda.is_available() and args.cuda:
print('Note: You are using GPU for training')
torch.cuda.set_device(args.gpu)
torch.cuda.manual_seed(args.seed)
if torch.cuda.is_available() and not args.cuda:
print('Warning: Using CPU for training')
dataset_map = {
'Reuters': Reuters,
'AAPD': AAPD,
'IMDB': IMDB,
'Yelp2014': Yelp2014
}
if args.dataset not in dataset_map:
raise ValueError('Unrecognized dataset')
else:
dataset_class = dataset_map[args.dataset]
train_iter, dev_iter, test_iter = dataset_class.iters(
args.data_dir,
args.word_vectors_file,
args.word_vectors_dir,
batch_size=args.batch_size,
device=args.gpu,
unk_init=UnknownWordVecCache.unk)
config = deepcopy(args)
config.dataset = train_iter.dataset
config.target_class = train_iter.dataset.NUM_CLASSES
print('Dataset:', args.dataset)
print('No. of target classes:', train_iter.dataset.NUM_CLASSES)
print('No. of train instances', len(train_iter.dataset))
print('No. of dev instances', len(dev_iter.dataset))
print('No. of test instances', len(test_iter.dataset))
if args.resume_snapshot:
if args.cuda:
model = torch.load(args.resume_snapshot,
map_location=lambda storage, location:
storage.cuda(args.gpu))
else:
model = torch.load(
args.resume_snapshot,
map_location=lambda storage, location: storage)
else:
model = CharCNN(config)
if args.cuda:
model.cuda()
if not args.trained_model:
save_path = os.path.join(args.save_path,
dataset_map[args.dataset].NAME)
os.makedirs(save_path, exist_ok=True)
parameter = filter(lambda p: p.requires_grad, model.parameters())
optimizer = torch.optim.Adam(parameter,
lr=args.lr,
weight_decay=args.weight_decay)
train_evaluator = EvaluatorFactory.get_evaluator(
dataset_class, model, None, train_iter, args.batch_size, args.gpu)
test_evaluator = EvaluatorFactory.get_evaluator(
dataset_class, model, None, test_iter, args.batch_size, args.gpu)
dev_evaluator = EvaluatorFactory.get_evaluator(dataset_class, model,
None, dev_iter,
args.batch_size,
args.gpu)
if hasattr(train_evaluator, 'is_multilabel'):
train_evaluator.is_multilabel = dataset_class.IS_MULTILABEL
if hasattr(dev_evaluator, 'is_multilabel'):
dev_evaluator.is_multilabel = dataset_class.IS_MULTILABEL
if hasattr(dev_evaluator, 'ignore_lengths'):
dev_evaluator.ignore_lengths = True
if hasattr(test_evaluator, 'is_multilabel'):
test_evaluator.is_multilabel = dataset_class.IS_MULTILABEL
if hasattr(test_evaluator, 'ignore_lengths'):
test_evaluator.ignore_lengths = True
trainer_config = {
'optimizer': optimizer,
'batch_size': args.batch_size,
'log_interval': args.log_every,
'patience': args.patience,
'model_outfile': args.save_path,
'logger': logger,
'is_multilabel': dataset_class.IS_MULTILABEL,
'ignore_lengths': True
}
trainer = TrainerFactory.get_trainer(args.dataset, model, None,
train_iter, trainer_config,
train_evaluator, test_evaluator,
dev_evaluator)
if not args.trained_model:
trainer.train(args.epochs)
else:
if args.cuda:
model = torch.load(args.trained_model,
map_location=lambda storage, location:
storage.cuda(args.gpu))
else:
model = torch.load(
args.trained_model,
map_location=lambda storage, location: storage)
# Calculate dev and test metrics
if hasattr(trainer, 'snapshot_path'):
model = torch.load(trainer.snapshot_path)
# evaluate_dataset(split_name, dataset_cls, model, embedding, loader, batch_size, device, is_multilabel
self.evaluate_dataset('dev', dataset_map[args.dataset], model, None,
dev_iter, args.batch_size, args.gpu,
dataset_class.IS_MULTILABEL)
self.evaluate_dataset('test', dataset_map[args.dataset], model, None,
test_iter, args.batch_size, args.gpu,
dataset_class.IS_MULTILABEL)
args.device)
dev_evaluator = EvaluatorFactory.get_evaluator(dataset_cls, model, embedding, dev_loader, args.batch_size,
args.device)
trainer_config = {
'optimizer': optimizer,
'batch_size': args.batch_size,
'log_interval': args.log_interval,
'model_outfile': args.model_outfile,
'lr_reduce_factor': args.lr_reduce_factor,
'patience': args.patience,
'tensorboard': args.tensorboard,
'run_label': args.run_label,
'logger': logger
}
trainer = TrainerFactory.get_trainer(args.dataset, model, embedding, train_loader, trainer_config, train_evaluator, test_evaluator, dev_evaluator)
if not args.skip_training:
total_params = 0
for param in model.parameters():
size = [s for s in param.size()]
total_params += np.prod(size)
logger.info('Total number of parameters: %s', total_params)
trainer.train(args.epochs)
_, _, state_dict, _, _ = load_checkpoint(args.model_outfile)
for k, tensor in state_dict.items():
state_dict[k] = tensor.to(device)
model.load_state_dict(state_dict)
test_evaluator.is_multilabel = dataset_class.IS_MULTILABEL
if hasattr(dev_evaluator, 'is_multilabel'):
dev_evaluator.is_multilabel = dataset_class.IS_MULTILABEL
trainer_config = {
'optimizer': optimizer,
'batch_size': args.batch_size,
'log_interval': args.log_every,
'patience': args.patience,
'model_outfile': args.save_path,
'logger': logger,
'is_multilabel': dataset_class.IS_MULTILABEL
}
trainer = TrainerFactory.get_trainer(args.dataset, model, None, train_iter,
trainer_config, train_evaluator,
test_evaluator, dev_evaluator)
if not args.trained_model:
trainer.train(args.epochs)
else:
if args.cuda:
model = torch.load(
args.trained_model,
map_location=lambda storage, location: storage.cuda(args.gpu))
else:
model = torch.load(args.trained_model,
map_location=lambda storage, location: storage)
# Calculate dev and test metrics
if hasattr(trainer, 'snapshot_path'):
dev_evaluator = EvaluatorFactory.get_evaluator(dataset_cls, model, embedding, dev_loader, args.batch_size, args.device)
trainer_config = {
'optimizer': optimizer,
'batch_size': args.batch_size,
'log_interval': args.log_interval,
'model_outfile': args.model_outfile,
'lr_reduce_factor': args.lr_reduce_factor,
'patience': args.patience,
'tensorboard': args.tensorboard,
'run_label': args.run_label,
'logger': logger,
'clip_norm': args.clip_norm
}
trainer = TrainerFactory.get_trainer('sick', model, embedding, train_loader, trainer_config, train_evaluator, test_evaluator, dev_evaluator)
if not args.skip_training:
total_params = 0
#print(model.parameters)
for param in model.parameters():
size = [s for s in param.size()]
total_params += np.prod(size)
#print(param.size())
#print('\n')
logger.info('Total number of parameters: %s', total_params)
trainer.train(args.epochs)
_, _, state_dict, _, _ = load_checkpoint(args.model_outfile)
for k, tensor in state_dict.items():