def create_model(args, vocab, embedding=None):
if args.use_glove:
embedding = get_glove_embedding(args.embedding_name, args.hidden_size,
vocab)
else:
embedding = None
logging.info('Creating IQ model...')
vqg = IQ(len(vocab),
args.max_length,
args.hidden_size,
args.num_categories,
vocab(vocab.SYM_SOQ),
vocab(vocab.SYM_EOS),
num_layers=args.num_layers,
rnn_cell=args.rnn_cell,
dropout_p=args.dropout_p,
input_dropout_p=args.input_dropout_p,
encoder_max_len=args.encoder_max_len,
embedding=embedding,
num_att_layers=args.num_att_layers,
z_size=args.z_size,
z_img=args.z_img,
z_category=args.z_category,
no_image_recon=args.no_image_recon,
no_category_space=args.no_category_space,
bayes=args.bayes)
return vqg
def create_model(args, vocab, embedding=None):
"""Creates the model.
Args:
args: Instance of Argument Parser.
vocab: Instance of Vocabulary.
Returns:
A VQG model.
"""
# Load GloVe embedding.
if args.use_glove:
embedding = get_glove_embedding(args.embedding_name,
300,vocab)
elif args.use_w2v:
embedding = get_wv_embedding(args.embedding_name,
200,vocab)
else:
embedding = None
# Build the models
logging.info('Creating IQ model...')
vqg = VQG(len(vocab), args.max_length, args.hidden_size,
vocab(vocab.SYM_SOQ), vocab(vocab.SYM_EOS),
num_layers=args.num_layers,
rnn_cell=args.rnn_cell,
dropout_p=args.dropout_p,
input_dropout_p=args.input_dropout_p,
encoder_max_len=args.encoder_max_len,
embedding=embedding,
num_att_layers=args.num_att_layers,
z_size=args.z_size)
return vqg
def create_model(args, vocab, embedding=None):
"""Creates the model.
Args:
args: Instance of Argument Parser.
vocab: Instance of Vocabulary.
Returns:
A multi class classification model.
"""
# Load GloVe embedding.
if args.use_glove:
embedding = get_glove_embedding(args.embedding_name, 300, vocab)
else:
embedding = None
# Build the models
logging.info('Creating multi-class classification model...')
model = Classifier(len(vocab),
embedding_dim=args.embedding_dim,
embedding=embedding,
hidden_dim=args.num_hidden_nodes,
output_dim=args.num_output_nodes,
num_layers=args.num_layers,
bidirectional=args.bidirectional,
dropout=args.dropout,
rnn_cell=args.rnn_cell)
return model
def main(args):
# Load the arguments.
model_dir = os.path.dirname(args.model_path)
params = Dict2Obj(
json.load(open(os.path.join(model_dir, "args.json"), "r")))
# Config logging
log_format = '%(levelname)-8s %(message)s'
logfile = os.path.join(model_dir, 'eval.log')
logging.basicConfig(filename=logfile,
level=logging.INFO,
format=log_format)
logging.getLogger().addHandler(logging.StreamHandler())
logging.info(json.dumps(args.__dict__))
# Load vocabulary wrapper.
vocab = load_vocab(params.vocab_path)
# Build data loader
logging.info("Building data loader...")
# Load GloVe embedding.
if params.use_glove:
embedding = get_glove_embedding(params.embedding_name, 300, vocab)
else:
embedding = None
# Processing input text
logging.info("Processing input text...")
text, length = process_text(args.text, vocab, max_length=20)
d_text = text
logging.info("Done")
# Build the models
logging.info('Creating IQ model...')
model = Classifier(len(vocab),
embedding_dim=params.embedding_dim,
embedding=embedding,
hidden_dim=params.num_hidden_nodes,
output_dim=params.num_output_nodes,
num_layers=params.num_layers,
bidirectional=params.bidirectional,
dropout=params.dropout,
rnn_cell=params.rnn_cell)
logging.info("Done")
logging.info("Loading model.")
model.load_state_dict(
torch.load(args.model_path + "model-tf-" + args.state + ".pkl"))
# Setup GPUs.
if torch.cuda.is_available():
logging.info("Using available GPU...")
model.cuda()
predict(model, d_text)
def create_model(args, vocab, embedding=None):
"""Creates the model.
Args:
args: Instance of Argument Parser.
vocab: Instance of Vocabulary.
Returns:
An IQ model.
"""
# Load GloVe embedding.
if args.use_glove:
embedding = get_glove_embedding(args.embedding_name, args.hidden_size,
vocab)
else:
embedding = None
# Build the models
logging.info('Creating IQ model...')
vqg = IQ(len(vocab),
args.max_length,
args.hidden_size,
args.num_categories,
vocab(vocab.SYM_SOQ),
vocab(vocab.SYM_EOS),
num_layers=args.num_layers,
rnn_cell=args.rnn_cell,
dropout_p=args.dropout_p,
input_dropout_p=args.input_dropout_p,
encoder_max_len=args.encoder_max_len,
embedding=embedding,
num_att_layers=args.num_att_layers,
use_attention=args.use_attention,
z_size=args.z_size,
no_answer_recon=args.no_answer_recon,
no_image_recon=args.no_image_recon,
no_category_space=args.no_category_space)
return vqg
def main(args):
"""Loads the model and then calls evaluate().
Args:
args: Instance of ArgumentParser.
"""
# Load the arguments.
model_dir = os.path.dirname(args.model_path)
params = Dict2Obj(
json.load(open(os.path.join(model_dir, "args.json"), "r")))
# Config logging
log_format = '%(levelname)-8s %(message)s'
logfile = os.path.join(model_dir, 'eval.log')
logging.basicConfig(filename=logfile,
level=logging.INFO,
format=log_format)
logging.getLogger().addHandler(logging.StreamHandler())
logging.info(json.dumps(args.__dict__))
# Image preprocessing
transform = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
# Load vocabulary wrapper.
vocab = load_vocab(params.vocab_path)
# Build data loader
logging.info("Building data loader...")
# Load GloVe embedding.
if params.use_glove:
embedding = get_glove_embedding(params.embedding_name, 300, vocab)
elif params.use_w2v:
embedding = get_wv_embedding(params.embedding_name, 200, vocab)
else:
embedding = None
# Build data loader
indices = []
ii = 0
for ii in range(500):
indices.append(ii)
logging.info("Building data loader...")
data_loader = get_loader(args.dataset,
transform,
args.batch_size,
shuffle=False,
num_workers=args.num_workers,
max_examples=args.max_examples,
indices=None)
logging.info("Done")
# Build the models
logging.info('Creating IQ model...')
vqa = VQA(
len(vocab),
params.max_length,
params.hidden_size,
vocab(vocab.SYM_SOQ),
vocab(vocab.SYM_EOS),
num_layers=params.num_layers,
rnn_cell=params.rnn_cell,
dropout_p=params.dropout_p,
input_dropout_p=params.input_dropout_p,
encoder_max_len=params.encoder_max_len,
embedding=embedding,
num_att_layers=params.num_att_layers,
#use_attention=params.use_attention,
z_size=params.z_size,
no_question_recon=params.no_question_recon,
no_image_recon=params.no_image_recon)
logging.info("Done")
logging.info("Loading model.")
vqa.load_state_dict(
torch.load(args.model_path + "vqa-tf-" + args.state + ".pkl"))
# Setup GPUs.
if torch.cuda.is_available():
logging.info("Using available GPU...")
vqa.cuda()
scores, gts, preds = evaluate(vqa, data_loader, vocab, args, params)
#gts, preds = evaluate(vqg, data_loader, vocab, args, params)
# Print and save the scores.
print(scores)
with open(os.path.join(model_dir, args.results_path), 'w') as results_file:
json.dump(scores, results_file)
with open(os.path.join(model_dir, args.preds_path), 'w') as preds_file:
json.dump(preds, preds_file)
with open(os.path.join(model_dir, args.gts_path), 'w') as gts_file:
json.dump(gts, gts_file)
def main(args):
# Load the arguments.
model_dir = os.path.dirname(args.model_path)
params = Dict2Obj(
json.load(open(os.path.join(model_dir, "args.json"), "r")))
# Config logging
log_format = '%(levelname)-8s %(message)s'
logfile = os.path.join(model_dir, 'eval.log')
logging.basicConfig(filename=logfile,
level=logging.INFO,
format=log_format)
logging.getLogger().addHandler(logging.StreamHandler())
logging.info(json.dumps(args.__dict__))
# Load vocabulary wrapper.
vocab = load_vocab(params.vocab_path)
# Build data loader
logging.info("Building data loader...")
# Load GloVe embedding.
if params.use_glove:
embedding = get_glove_embedding(params.embedding_name, 300, vocab)
else:
embedding = None
# Build data loader
logging.info("Building data loader...")
data_loader = get_loader(args.dataset,
args.batch_size,
shuffle=False,
num_workers=args.num_workers,
max_examples=args.max_examples)
logging.info("Done")
# Build the models
logging.info('Creating a multi class classification model...')
model = Classifier(len(vocab),
embedding_dim=params.embedding_dim,
embedding=embedding,
hidden_dim=params.num_hidden_nodes,
output_dim=params.num_output_nodes,
num_layers=params.num_layers,
bidirectional=params.bidirectional,
dropout=params.dropout,
rnn_cell=params.rnn_cell)
logging.info("Done")
logging.info("Loading model.")
model.load_state_dict(
torch.load(args.model_path + "model-tf-" + args.state + ".pkl"))
# Setup GPUs.
if torch.cuda.is_available():
logging.info("Using available GPU...")
model.cuda()
scores, gts, preds = evaluate(model, data_loader, vocab, args, params)
# Print and save the scores.
print(scores)
with open(os.path.join(model_dir, args.results_path), 'w') as results_file:
json.dump(scores, results_file)
with open(os.path.join(model_dir, args.preds_path), 'w') as preds_file:
json.dump(preds, preds_file)
with open(os.path.join(model_dir, args.gts_path), 'w') as gts_file:
json.dump(gts, gts_file)