def main() -> None:
tokenizer = Tokenizer(args.vocab_file)
vocabulary_size = len(tokenizer)
dataset = SentenceDataset(args.input_file, tokenizer=tokenizer.encode)
loader = DataLoader(dataset,
args.batch_size,
shuffle=False,
collate_fn=dataset.collate_fn,
drop_last=False)
searcher = BeamSearch(tokenizer.eos_index, beam_size=args.search_width)
model = VAE(
num_embeddings=len(tokenizer),
dim_embedding=args.dim_embedding,
dim_hidden=args.dim_hidden,
dim_latent=args.dim_latent,
num_layers=args.num_layers,
bidirectional=args.bidirectional,
dropout=0.,
word_dropout=0.,
dropped_index=tokenizer.unk_index,
).to(device)
model.load_state_dict(torch.load(args.checkpoint_file,
map_location=device))
model.eval()
print('Generating sentence...')
all_hypotheses = []
with torch.no_grad():
for s in tqdm(loader):
s = s.to(device)
length = torch.sum(s != tokenizer.pad_index, dim=-1)
bsz = s.shape[0]
mean, logvar = model.encode(s, length)
# z = model.reparameterize(mean, logvar)
z = mean
hidden = model.fc_hidden(z)
hidden = hidden.view(bsz, -1,
model.dim_hidden).transpose(0,
1).contiguous()
start_predictions = torch.zeros(bsz, device=device).fill_(
tokenizer.bos_index).long()
start_state = {'hidden': hidden.permute(1, 0, 2)}
predictions, log_probabilities = searcher.search(
start_predictions, start_state, model.step)
for preds in predictions:
tokens = preds[0]
tokens = tokens[tokens != tokenizer.eos_index].tolist()
all_hypotheses.append(tokenizer.decode(tokens))
print('Done')
with open(args.output_file, 'w') as f:
f.write('\n'.join(all_hypotheses))
def main() -> None:
tokenizer = Tokenizer(args.vocab_file)
vocabulary_size = len(tokenizer)
searcher = BeamSearch(tokenizer.eos_index, beam_size=args.search_width)
model = VAE(
num_embeddings=len(tokenizer),
dim_embedding=args.dim_embedding,
dim_hidden=args.dim_hidden,
dim_latent=args.dim_latent,
num_layers=args.num_layers,
bidirectional=args.bidirectional,
dropout=0.,
word_dropout=0.,
dropped_index=tokenizer.unk_index,
).to(device)
model.load_state_dict(torch.load(args.checkpoint_file,
map_location=device))
model.eval()
sentence1 = input('Please input sentence1: ')
sentence2 = input('Please input sentence2: ')
s1 = [tokenizer.bos_index
] + tokenizer.encode(sentence1) + [tokenizer.eos_index]
s2 = [tokenizer.bos_index
] + tokenizer.encode(sentence2) + [tokenizer.eos_index]
z1, _ = model.encode(
torch.tensor([s1]).to(device),
torch.tensor([len(s1)]).to(device))
z2, _ = model.encode(
torch.tensor([s2]).to(device),
torch.tensor([len(s2)]).to(device))
print("\nGenerate intermediate sentences")
print(" %s" % sentence1)
for r in range(1, 10):
z = (1 - 0.1 * r) * z1 + 0.1 * r * z2
hidden = model.fc_hidden(z)
hidden = hidden.view(1, -1,
model.dim_hidden).transpose(0, 1).contiguous()
start_predictions = torch.zeros(1, device=device).fill_(
tokenizer.bos_index).long()
start_state = {'hidden': hidden.permute(1, 0, 2)}
predictions, log_probabilities = searcher.search(
start_predictions, start_state, model.step)
tokens = predictions[0, 0]
tokens = tokens[tokens != tokenizer.eos_index].tolist()
print("[%d:%d] %s" % (10 - r, r, tokenizer.decode(tokens)))
print(" %s" % sentence2)
def main() -> None:
tokenizer = Tokenizer(args.vocab_file)
vocabulary_size = len(tokenizer)
searcher = BeamSearch(tokenizer.eos_index, beam_size=args.search_width)
model = VAE(
num_embeddings=len(tokenizer),
dim_embedding=args.dim_embedding,
dim_hidden=args.dim_hidden,
dim_latent=args.dim_latent,
num_layers=args.num_layers,
bidirectional=args.bidirectional,
dropout=0.,
word_dropout=0.,
dropped_index=tokenizer.unk_index,
).to(device)
model.load_state_dict(torch.load(args.checkpoint_file,
map_location=device))
model.eval()
z = torch.randn(args.sample_size, args.dim_latent, device=device)
hidden = model.fc_hidden(z)
hidden = hidden.view(args.sample_size, -1,
model.dim_hidden).transpose(0, 1).contiguous()
start_predictions = torch.zeros(args.sample_size, device=device).fill_(
tokenizer.bos_index).long()
start_state = {'hidden': hidden.permute(1, 0, 2)}
predictions, log_probabilities = searcher.search(start_predictions,
start_state, model.step)
for pred in predictions:
tokens = pred[0]
tokens = tokens[tokens != tokenizer.eos_index].tolist()
print(tokenizer.decode(tokens))
