def sample(args):
print('loading VAE at %s' % args.load_model, file=sys.stderr)
fname, ext = os.path.splitext(args.load_model)
encoder_path = fname + '.encoder' + ext
decoder_path = fname + '.decoder' + ext
vae_params = torch.load(args.load_model, map_location=lambda storage, loc: storage)
encoder_params = torch.load(encoder_path, map_location=lambda storage, loc: storage)
decoder_params = torch.load(decoder_path, map_location=lambda storage, loc: storage)
transition_system = encoder_params['transition_system']
vae_params['args'].cuda = encoder_params['args'].cuda = decoder_params['args'].cuda = args.cuda
decoder = Reconstructor(decoder_params['args'], decoder_params['vocab'], transition_system)
decoder.load_state_dict(decoder_params['state_dict'])
assert vae_params['args'].prior == 'lstm'
prior = LSTMPrior.load(args.load_prior, transition_system=decoder_params['transition_system'], cuda=args.cuda)
print('loaded prior at %s' % args.load_prior, file=sys.stderr)
# freeze prior parameters
for p in prior.parameters():
p.requires_grad = False
decoder.eval()
prior.eval()
if args.cuda:
decoder.cuda()
prior.cuda()
err_num = 0
total_num = 0
# while True:
for sample_id in xrange(10000):
sampled_z = prior.sample()
sampled_z = ' '.join(sampled_z)
sampled_z = sampled_z.replace(' else :', 'else :').replace(' except ', 'except ').replace(' elif ', 'elif ').replace('<unk>', 'unk')
print('Z: %s' % sampled_z)
total_num += 1
try:
transition_system.surface_code_to_ast(sampled_z)
except:
print('Error!')
err_num += 1
continue
print('Sampled NL sentences:')
sampled_nls = decoder.sample(sampled_z)
for i, sampled_nl in enumerate(sampled_nls):
print('[%d] %s' % (i, ' '.join(sampled_nl)))
print()
print('Ratio of well-formed samples: %d/%d=%.5f' % (total_num - err_num, total_num,
(total_num - err_num) / float(total_num)), file=sys.stderr)
def train_decoder(args):
train_set = Dataset.from_bin_file(args.train_file)
dev_set = Dataset.from_bin_file(args.dev_file)
vocab = pickle.load(open(args.vocab))
grammar = ASDLGrammar.from_text(open(args.asdl_file).read())
transition_system = TransitionSystem.get_class_by_lang(args.lang)(grammar)
model = Reconstructor(args, vocab, transition_system)
model.train()
if args.cuda: model.cuda()
optimizer = torch.optim.Adam(model.parameters(), lr=args.lr)
def evaluate_ppl():
model.eval()
cum_loss = 0.
cum_tgt_words = 0.
for batch in dev_set.batch_iter(args.batch_size):
loss = -model.score(batch).sum()
cum_loss += loss.data[0]
cum_tgt_words += sum(len(e.src_sent) + 1
for e in batch) # add ending </s>
ppl = np.exp(cum_loss / cum_tgt_words)
model.train()
return ppl
print('begin training decoder, %d training examples, %d dev examples' %
(len(train_set), len(dev_set)),
file=sys.stderr)
print('vocab: %s' % repr(vocab), file=sys.stderr)
epoch = train_iter = 0
report_loss = report_examples = 0.
history_dev_scores = []
num_trial = patience = 0
while True:
epoch += 1
epoch_begin = time.time()
for batch_examples in train_set.batch_iter(batch_size=args.batch_size,
shuffle=True):
batch_examples = [
e for e in batch_examples
if len(e.tgt_actions) <= args.decode_max_time_step
]
# batch_examples = [e for e in train_set.examples if e.idx in [10192, 10894, 9706, 4659, 5609, 1442, 5849, 10644, 4592, 1875]]
train_iter += 1
optimizer.zero_grad()
loss = -model.score(batch_examples)
# print(loss.data)
loss_val = torch.sum(loss).data[0]
report_loss += loss_val
report_examples += len(batch_examples)
loss = torch.mean(loss)
loss.backward()
# clip gradient
grad_norm = torch.nn.utils.clip_grad_norm(model.parameters(),
args.clip_grad)
optimizer.step()
if train_iter % args.log_every == 0:
print('[Iter %d] encoder loss=%.5f' %
(train_iter, report_loss / report_examples),
file=sys.stderr)
report_loss = report_examples = 0.
print('[Epoch %d] epoch elapsed %ds' %
(epoch, time.time() - epoch_begin),
file=sys.stderr)
# model_file = args.save_to + '.iter%d.bin' % train_iter
# print('save model to [%s]' % model_file, file=sys.stderr)
# model.save(model_file)
# perform validation
print('[Epoch %d] begin validation' % epoch, file=sys.stderr)
eval_start = time.time()
# evaluate ppl
ppl = evaluate_ppl()
print('[Epoch %d] ppl=%.5f took %ds' %
(epoch, ppl, time.time() - eval_start),
file=sys.stderr)
dev_acc = -ppl
is_better = history_dev_scores == [] or dev_acc > max(
history_dev_scores)
history_dev_scores.append(dev_acc)
if is_better:
patience = 0
model_file = args.save_to + '.bin'
print('save currently the best model ..', file=sys.stderr)
print('save model to [%s]' % model_file, file=sys.stderr)
model.save(model_file)
# also save the optimizers' state
torch.save(optimizer.state_dict(), args.save_to + '.optim.bin')
elif patience < args.patience:
patience += 1
print('hit patience %d' % patience, file=sys.stderr)
if patience == args.patience:
num_trial += 1
print('hit #%d trial' % num_trial, file=sys.stderr)
if num_trial == args.max_num_trial:
print('early stop!', file=sys.stderr)
exit(0)
# decay lr, and restore from previously best checkpoint
lr = optimizer.param_groups[0]['lr'] * args.lr_decay
print('load previously best model and decay learning rate to %f' %
lr,
file=sys.stderr)
# load model
params = torch.load(args.save_to + '.bin',
map_location=lambda storage, loc: storage)
model.load_state_dict(params['state_dict'])
if args.cuda: model = model.cuda()
# load optimizers
if args.reset_optimizer:
print('reset optimizer', file=sys.stderr)
optimizer = torch.optim.Adam(
model.inference_model.parameters(), lr=lr)
else:
print('restore parameters of the optimizers', file=sys.stderr)
optimizer.load_state_dict(
torch.load(args.save_to + '.optim.bin'))
# set new lr
for param_group in optimizer.param_groups:
param_group['lr'] = lr
# reset patience
patience = 0
def train_decoder(args):
train_set = Dataset.from_bin_file(args.train_file)
dev_set = Dataset.from_bin_file(args.dev_file)
vocab = pickle.load(open(args.vocab))
grammar = ASDLGrammar.from_text(open(args.asdl_file).read())
transition_system = TransitionSystem.get_class_by_lang(args.lang)(grammar)
model = Reconstructor(args, vocab, transition_system)
model.train()
if args.cuda: model.cuda()
optimizer = torch.optim.Adam(model.parameters(), lr=args.lr)
def evaluate_ppl():
model.eval()
cum_loss = 0.
cum_tgt_words = 0.
for batch in dev_set.batch_iter(args.batch_size):
loss = -model.score(batch).sum()
cum_loss += loss.data[0]
cum_tgt_words += sum(len(e.src_sent) + 1 for e in batch) # add ending </s>
ppl = np.exp(cum_loss / cum_tgt_words)
model.train()
return ppl
print('begin training decoder, %d training examples, %d dev examples' % (len(train_set), len(dev_set)), file=sys.stderr)
print('vocab: %s' % repr(vocab), file=sys.stderr)
epoch = train_iter = 0
report_loss = report_examples = 0.
history_dev_scores = []
num_trial = patience = 0
while True:
epoch += 1
epoch_begin = time.time()
for batch_examples in train_set.batch_iter(batch_size=args.batch_size, shuffle=True):
batch_examples = [e for e in batch_examples if len(e.tgt_actions) <= args.decode_max_time_step]
# batch_examples = [e for e in train_set.examples if e.idx in [10192, 10894, 9706, 4659, 5609, 1442, 5849, 10644, 4592, 1875]]
train_iter += 1
optimizer.zero_grad()
loss = -model.score(batch_examples)
# print(loss.data)
loss_val = torch.sum(loss).data[0]
report_loss += loss_val
report_examples += len(batch_examples)
loss = torch.mean(loss)
loss.backward()
# clip gradient
grad_norm = torch.nn.utils.clip_grad_norm(model.parameters(), args.clip_grad)
optimizer.step()
if train_iter % args.log_every == 0:
print('[Iter %d] encoder loss=%.5f' %
(train_iter,
report_loss / report_examples),
file=sys.stderr)
report_loss = report_examples = 0.
print('[Epoch %d] epoch elapsed %ds' % (epoch, time.time() - epoch_begin), file=sys.stderr)
# model_file = args.save_to + '.iter%d.bin' % train_iter
# print('save model to [%s]' % model_file, file=sys.stderr)
# model.save(model_file)
# perform validation
print('[Epoch %d] begin validation' % epoch, file=sys.stderr)
eval_start = time.time()
# evaluate ppl
ppl = evaluate_ppl()
print('[Epoch %d] ppl=%.5f took %ds' % (epoch, ppl, time.time() - eval_start), file=sys.stderr)
dev_acc = -ppl
is_better = history_dev_scores == [] or dev_acc > max(history_dev_scores)
history_dev_scores.append(dev_acc)
if is_better:
patience = 0
model_file = args.save_to + '.bin'
print('save currently the best model ..', file=sys.stderr)
print('save model to [%s]' % model_file, file=sys.stderr)
model.save(model_file)
# also save the optimizers' state
torch.save(optimizer.state_dict(), args.save_to + '.optim.bin')
elif patience < args.patience:
patience += 1
print('hit patience %d' % patience, file=sys.stderr)
if patience == args.patience:
num_trial += 1
print('hit #%d trial' % num_trial, file=sys.stderr)
if num_trial == args.max_num_trial:
print('early stop!', file=sys.stderr)
exit(0)
# decay lr, and restore from previously best checkpoint
lr = optimizer.param_groups[0]['lr'] * args.lr_decay
print('load previously best model and decay learning rate to %f' % lr, file=sys.stderr)
# load model
params = torch.load(args.save_to + '.bin', map_location=lambda storage, loc: storage)
model.load_state_dict(params['state_dict'])
if args.cuda: model = model.cuda()
# load optimizers
if args.reset_optimizer:
print('reset optimizer', file=sys.stderr)
optimizer = torch.optim.Adam(model.inference_model.parameters(), lr=lr)
else:
print('restore parameters of the optimizers', file=sys.stderr)
optimizer.load_state_dict(torch.load(args.save_to + '.optim.bin'))
# set new lr
for param_group in optimizer.param_groups:
param_group['lr'] = lr
# reset patience
patience = 0
def log_semi(args):
print('loading VAE at %s' % args.load_model, file=sys.stderr)
fname, ext = os.path.splitext(args.load_model)
encoder_path = fname + '.encoder' + ext
decoder_path = fname + '.decoder' + ext
vae_params = torch.load(args.load_model, map_location=lambda storage, loc: storage)
encoder_params = torch.load(encoder_path, map_location=lambda storage, loc: storage)
decoder_params = torch.load(decoder_path, map_location=lambda storage, loc: storage)
transition_system = encoder_params['transition_system']
vae_params['args'].cuda = encoder_params['args'].cuda = decoder_params['args'].cuda = args.cuda
encoder = Parser(encoder_params['args'], encoder_params['vocab'], transition_system)
decoder = Reconstructor(decoder_params['args'], decoder_params['vocab'], transition_system)
if vae_params['args'].prior == 'lstm':
prior = LSTMPrior.load(vae_params['args'].load_prior, transition_system=decoder_params['transition_system'], cuda=args.cuda)
print('loaded prior at %s' % vae_params['args'].load_prior, file=sys.stderr)
# freeze prior parameters
for p in prior.parameters():
p.requires_grad = False
prior.eval()
else:
prior = UniformPrior()
if vae_params['args'].baseline == 'mlp':
structVAE = StructVAE(encoder, decoder, prior, vae_params['args'])
elif vae_params['args'].baseline == 'src_lm' or vae_params['args'].baseline == 'src_lm_and_linear':
src_lm = LSTMLanguageModel.load(vae_params['args'].load_src_lm)
print('loaded source LM at %s' % vae_params['args'].load_src_lm, file=sys.stderr)
Baseline = StructVAE_LMBaseline if args.baseline == 'src_lm' else StructVAE_SrcLmAndLinearBaseline
structVAE = Baseline(encoder, decoder, prior, src_lm, vae_params['args'])
else:
raise ValueError('unknown baseline')
structVAE.load_parameters(args.load_model)
structVAE.train()
if args.cuda: structVAE.cuda()
unlabeled_data = Dataset.from_bin_file(args.unlabeled_file) # pretend they are un-labeled!
print('*** begin sampling ***', file=sys.stderr)
start_time = time.time()
train_iter = 0
log_entries = []
for unlabeled_examples in unlabeled_data.batch_iter(batch_size=args.batch_size, shuffle=False):
unlabeled_examples = [e for e in unlabeled_examples if len(e.tgt_actions) <= args.decode_max_time_step]
train_iter += 1
try:
unsup_encoder_loss, unsup_decoder_loss, unsup_baseline_loss, meta_data = structVAE.get_unsupervised_loss(
unlabeled_examples)
except ValueError as e:
print(e.message, file=sys.stderr)
continue
samples = meta_data['samples']
for v in meta_data.itervalues():
if isinstance(v, Variable): v.cpu()
for i, sample in enumerate(samples):
ref_example = [e for e in unlabeled_examples if e.idx == int(sample.idx[:sample.idx.index('-')])][0]
log_entry = {
'sample': sample,
'ref_example': ref_example,
'log_p_z_x': meta_data['encoding_scores'][i].data[0],
'log_p_x_z': meta_data['reconstruction_scores'][i].data[0],
'kl': meta_data['kl_term'][i].data[0],
'prior': meta_data['prior'][i].data[0],
'baseline': meta_data['baseline'][i].data[0],
'learning_signal': meta_data['raw_learning_signal'][i].data[0],
'learning_signal - baseline': meta_data['learning_signal'][i].data[0],
'encoder_loss': unsup_encoder_loss[i].data[0],
'decoder_loss': unsup_decoder_loss[i].data[0]
}
log_entries.append(log_entry)
print('done! took %d s' % (time.time() - start_time), file=sys.stderr)
pkl.dump(log_entries, open(args.save_to, 'wb'))
def train_semi(args):
encoder_params = torch.load(args.load_model, map_location=lambda storage, loc: storage)
decoder_params = torch.load(args.load_decoder, map_location=lambda storage, loc: storage)
print('loaded encoder at %s' % args.load_model, file=sys.stderr)
print('loaded decoder at %s' % args.load_decoder, file=sys.stderr)
transition_system = encoder_params['transition_system']
encoder_params['args'].cuda = decoder_params['args'].cuda = args.cuda
encoder = Parser(encoder_params['args'], encoder_params['vocab'], transition_system)
encoder.load_state_dict(encoder_params['state_dict'])
decoder = Reconstructor(decoder_params['args'], decoder_params['vocab'], transition_system)
decoder.load_state_dict(decoder_params['state_dict'])
if args.prior == 'lstm':
prior = LSTMPrior.load(args.load_prior, transition_system=transition_system, cuda=args.cuda)
print('loaded prior at %s' % args.load_prior, file=sys.stderr)
# freeze prior parameters
for p in prior.parameters():
p.requires_grad = False
prior.eval()
else:
prior = UniformPrior()
if args.baseline == 'mlp':
structVAE = StructVAE(encoder, decoder, prior, args)
elif args.baseline == 'src_lm' or args.baseline == 'src_lm_and_linear':
src_lm = LSTMLanguageModel.load(args.load_src_lm)
print('loaded source LM at %s' % args.load_src_lm, file=sys.stderr)
vae_cls = StructVAE_LMBaseline if args.baseline == 'src_lm' else StructVAE_SrcLmAndLinearBaseline
structVAE = vae_cls(encoder, decoder, prior, src_lm, args)
else:
raise ValueError('unknown baseline')
structVAE.train()
if args.cuda: structVAE.cuda()
labeled_data = Dataset.from_bin_file(args.train_file)
# labeled_data.examples = labeled_data.examples[:10]
unlabeled_data = Dataset.from_bin_file(args.unlabeled_file) # pretend they are un-labeled!
dev_set = Dataset.from_bin_file(args.dev_file)
# dev_set.examples = dev_set.examples[:10]
optimizer = torch.optim.Adam(ifilter(lambda p: p.requires_grad, structVAE.parameters()), lr=args.lr)
print('*** begin semi-supervised training %d labeled examples, %d unlabeled examples ***' %
(len(labeled_data), len(unlabeled_data)), file=sys.stderr)
report_encoder_loss = report_decoder_loss = report_src_sent_words_num = report_tgt_query_words_num = report_examples = 0.
report_unsup_examples = report_unsup_encoder_loss = report_unsup_decoder_loss = report_unsup_baseline_loss = 0.
patience = 0
num_trial = 1
epoch = train_iter = 0
history_dev_scores = []
while True:
epoch += 1
epoch_begin = time.time()
unlabeled_examples_iter = unlabeled_data.batch_iter(batch_size=args.unsup_batch_size, shuffle=True)
for labeled_examples in labeled_data.batch_iter(batch_size=args.batch_size, shuffle=True):
labeled_examples = [e for e in labeled_examples if len(e.tgt_actions) <= args.decode_max_time_step]
train_iter += 1
optimizer.zero_grad()
report_examples += len(labeled_examples)
sup_encoder_loss = -encoder.score(labeled_examples)
sup_decoder_loss = -decoder.score(labeled_examples)
report_encoder_loss += sup_encoder_loss.sum().data[0]
report_decoder_loss += sup_decoder_loss.sum().data[0]
sup_encoder_loss = torch.mean(sup_encoder_loss)
sup_decoder_loss = torch.mean(sup_decoder_loss)
sup_loss = sup_encoder_loss + sup_decoder_loss
# compute unsupervised loss
try:
unlabeled_examples = next(unlabeled_examples_iter)
except StopIteration:
# if finished unlabeled data stream, restart it
unlabeled_examples_iter = unlabeled_data.batch_iter(batch_size=args.batch_size, shuffle=True)
unlabeled_examples = next(unlabeled_examples_iter)
unlabeled_examples = [e for e in unlabeled_examples if len(e.tgt_actions) <= args.decode_max_time_step]
try:
unsup_encoder_loss, unsup_decoder_loss, unsup_baseline_loss, meta_data = structVAE.get_unsupervised_loss(
unlabeled_examples)
nan = False
if nn_utils.isnan(sup_loss.data):
print('Nan in sup_loss')
nan = True
if nn_utils.isnan(unsup_encoder_loss.data):
print('Nan in unsup_encoder_loss!', file=sys.stderr)
nan = True
if nn_utils.isnan(unsup_decoder_loss.data):
print('Nan in unsup_decoder_loss!', file=sys.stderr)
nan = True
if nn_utils.isnan(unsup_baseline_loss.data):
print('Nan in unsup_baseline_loss!', file=sys.stderr)
nan = True
if nan:
# torch.save((unsup_encoder_loss, unsup_decoder_loss, unsup_baseline_loss, meta_data), 'nan_data.bin')
continue
report_unsup_encoder_loss += unsup_encoder_loss.sum().data[0]
report_unsup_decoder_loss += unsup_decoder_loss.sum().data[0]
report_unsup_baseline_loss += unsup_baseline_loss.sum().data[0]
report_unsup_examples += unsup_encoder_loss.size(0)
except ValueError as e:
print(e.message, file=sys.stderr)
continue
# except Exception as e:
# print('********** Error **********', file=sys.stderr)
# print('batch labeled examples: ', file=sys.stderr)
# for example in labeled_examples:
# print('%s %s' % (example.idx, ' '.join(example.src_sent)), file=sys.stderr)
# print('batch unlabeled examples: ', file=sys.stderr)
# for example in unlabeled_examples:
# print('%s %s' % (example.idx, ' '.join(example.src_sent)), file=sys.stderr)
# print(e.message, file=sys.stderr)
# traceback.print_exc(file=sys.stderr)
# for k, v in meta_data.iteritems():
# print('%s: %s' % (k, v), file=sys.stderr)
# print('********** Error **********', file=sys.stderr)
# continue
unsup_loss = torch.mean(unsup_encoder_loss) + torch.mean(unsup_decoder_loss) + torch.mean(unsup_baseline_loss)
loss = sup_loss + args.unsup_loss_weight * unsup_loss
loss.backward()
# clip gradient
grad_norm = torch.nn.utils.clip_grad_norm(structVAE.parameters(), args.clip_grad)
optimizer.step()
if train_iter % args.log_every == 0:
print('[Iter %d] supervised: encoder loss=%.5f, decoder loss=%.5f' %
(train_iter,
report_encoder_loss / report_examples,
report_decoder_loss / report_examples),
file=sys.stderr)
print('[Iter %d] unsupervised: encoder loss=%.5f, decoder loss=%.5f, baseline loss=%.5f' %
(train_iter,
report_unsup_encoder_loss / report_unsup_examples,
report_unsup_decoder_loss / report_unsup_examples,
report_unsup_baseline_loss / report_unsup_examples),
file=sys.stderr)
# print('[Iter %d] unsupervised: baseline=%.5f, raw learning signal=%.5f, learning signal=%.5f' % (train_iter,
# meta_data['baseline'].mean().data[0],
# meta_data['raw_learning_signal'].mean().data[0],
# meta_data['learning_signal'].mean().data[0]), file=sys.stderr)
if isinstance(structVAE, StructVAE_LMBaseline):
print('[Iter %d] baseline: source LM b_lm_weight: %.3f, b: %.3f' % (train_iter,
structVAE.b_lm_weight.data[0],
structVAE.b.data[0]),
file=sys.stderr)
samples = meta_data['samples']
for v in meta_data.itervalues():
if isinstance(v, Variable): v.cpu()
for i, sample in enumerate(samples[:15]):
print('\t[%s] Source: %s' % (sample.idx, ' '.join(sample.src_sent)), file=sys.stderr)
print('\t[%s] Code: \n%s' % (sample.idx, sample.tgt_code), file=sys.stderr)
ref_example = [e for e in unlabeled_examples if e.idx == int(sample.idx[:sample.idx.index('-')])][0]
print('\t[%s] Gold Code: \n%s' % (sample.idx, ref_example.tgt_code), file=sys.stderr)
print('\t[%s] Log p(z|x): %f' % (sample.idx, meta_data['encoding_scores'][i].data[0]), file=sys.stderr)
print('\t[%s] Log p(x|z): %f' % (sample.idx, meta_data['reconstruction_scores'][i].data[0]), file=sys.stderr)
print('\t[%s] KL term: %f' % (sample.idx, meta_data['kl_term'][i].data[0]), file=sys.stderr)
print('\t[%s] Prior: %f' % (sample.idx, meta_data['prior'][i].data[0]), file=sys.stderr)
print('\t[%s] baseline: %f' % (sample.idx, meta_data['baseline'][i].data[0]), file=sys.stderr)
print('\t[%s] Raw Learning Signal: %f' % (sample.idx, meta_data['raw_learning_signal'][i].data[0]), file=sys.stderr)
print('\t[%s] Learning Signal - baseline: %f' % (sample.idx, meta_data['learning_signal'][i].data[0]), file=sys.stderr)
print('\t[%s] Encoder Loss: %f' % (sample.idx, unsup_encoder_loss[i].data[0]), file=sys.stderr)
print('\t**************************', file=sys.stderr)
report_encoder_loss = report_decoder_loss = report_examples = 0.
report_unsup_encoder_loss = report_unsup_decoder_loss = report_unsup_baseline_loss = report_unsup_examples = 0.
print('[Epoch %d] epoch elapsed %ds' % (epoch, time.time() - epoch_begin), file=sys.stderr)
# perform validation
print('[Epoch %d] begin validation' % epoch, file=sys.stderr)
eval_start = time.time()
eval_results = evaluation.evaluate(dev_set.examples, encoder, args, verbose=True)
dev_acc = eval_results['accuracy']
print('[Epoch %d] code generation accuracy=%.5f took %ds' % (epoch, dev_acc, time.time() - eval_start),
file=sys.stderr)
is_better = history_dev_scores == [] or dev_acc > max(history_dev_scores)
history_dev_scores.append(dev_acc)
# model_file = args.save_to + '.iter%d.bin' % train_iter
# print('save model to [%s]' % model_file, file=sys.stderr)
# structVAE.save(model_file)
if is_better:
patience = 0
model_file = args.save_to + '.bin'
print('save currently the best model ..', file=sys.stderr)
print('save model to [%s]' % model_file, file=sys.stderr)
structVAE.save(model_file)
# also save the optimizers' state
torch.save(optimizer.state_dict(), args.save_to + '.optim.bin')
elif epoch == args.max_epoch:
print('reached max epoch, stop!', file=sys.stderr)
exit(0)
elif patience < args.patience:
patience += 1
print('hit patience %d' % patience, file=sys.stderr)
if patience == args.patience:
num_trial += 1
print('hit #%d trial' % num_trial, file=sys.stderr)
if num_trial == args.max_num_trial:
print('early stop!', file=sys.stderr)
exit(0)
# decay lr, and restore from previously best checkpoint
lr = optimizer.param_groups[0]['lr'] * args.lr_decay
print('load previously best model and decay learning rate to %f' % lr, file=sys.stderr)
# load best model's parameters
structVAE.load_parameters(args.save_to + '.bin')
if args.cuda: structVAE = structVAE.cuda()
# load optimizers
if args.reset_optimizer:
print('reset to a new infer_optimizer', file=sys.stderr)
optimizer = torch.optim.Adam(ifilter(lambda p: p.requires_grad, structVAE.parameters()), lr=lr)
else:
print('restore parameters of the optimizers', file=sys.stderr)
optimizer.load_state_dict(torch.load(args.save_to + '.optim.bin'))
# set new lr
for param_group in optimizer.param_groups:
param_group['lr'] = lr
# reset patience
patience = 0
def train_semi_jae(args):
bi_direction = args.bi_direction
encoder_params = torch.load(args.load_model,
map_location=lambda storage, loc: storage)
decoder_params = torch.load(args.load_decoder,
map_location=lambda storage, loc: storage)
print('loaded encoder at %s' % args.load_model, file=sys.stderr)
print('loaded decoder at %s' % args.load_decoder, file=sys.stderr)
transition_system = encoder_params['transition_system']
encoder_params['args'].cuda = decoder_params['args'].cuda = args.cuda
encoder = Parser(encoder_params['args'], encoder_params['vocab'],
transition_system)
encoder.load_state_dict(encoder_params['state_dict'])
decoder = Reconstructor(decoder_params['args'], decoder_params['vocab'],
transition_system)
decoder.load_state_dict(decoder_params['state_dict'])
zprior = LSTMPrior.load(args.load_prior,
transition_system=transition_system,
cuda=args.cuda)
print('loaded p(z) prior at %s' % args.load_prior, file=sys.stderr)
# freeze prior parameters
for p in zprior.parameters():
p.requires_grad = False
zprior.eval()
xprior = LSTMLanguageModel.load(args.load_src_lm)
print('loaded p(x) prior at %s' % args.load_src_lm, file=sys.stderr)
xprior.eval()
if args.cache:
jae = JAE_cache(encoder, decoder, zprior, xprior, args)
else:
jae = JAE(encoder, decoder, zprior, xprior, args)
jae.train()
encoder.train()
decoder.train()
if args.cuda: jae.cuda()
labeled_data = Dataset.from_bin_file(args.train_file)
# labeled_data.examples = labeled_data.examples[:10]
unlabeled_data = Dataset.from_bin_file(
args.unlabeled_file) # pretend they are un-labeled!
dev_set = Dataset.from_bin_file(args.dev_file)
# dev_set.examples = dev_set.examples[:10]
optimizer = torch.optim.Adam(
[p for p in jae.parameters() if p.requires_grad], lr=args.lr)
print(
'*** begin semi-supervised training %d labeled examples, %d unlabeled examples ***'
% (len(labeled_data), len(unlabeled_data)),
file=sys.stderr)
report_encoder_loss = report_decoder_loss = report_examples = 0.
report_unsup_examples = report_unsup_encoder_loss = report_unsup_decoder_loss = report_unsup_baseline_loss = 0.
patience = 0
num_trial = 1
epoch = train_iter = 0
history_dev_scores = []
while True:
epoch += 1
epoch_begin = time.time()
unlabeled_examples_iter = unlabeled_data.batch_iter(
batch_size=args.unsup_batch_size, shuffle=True)
for labeled_examples in labeled_data.batch_iter(
batch_size=args.batch_size, shuffle=True):
labeled_examples = [
e for e in labeled_examples
if len(e.tgt_actions) <= args.decode_max_time_step
]
train_iter += 1
optimizer.zero_grad()
report_examples += len(labeled_examples)
sup_encoder_loss = -encoder.score(labeled_examples)
sup_decoder_loss = -decoder.score(labeled_examples)
report_encoder_loss += sup_encoder_loss.sum().data[0]
report_decoder_loss += sup_decoder_loss.sum().data[0]
sup_encoder_loss = torch.mean(sup_encoder_loss)
sup_decoder_loss = torch.mean(sup_decoder_loss)
sup_loss = sup_encoder_loss + sup_decoder_loss
# compute unsupervised loss
try:
unlabeled_examples = next(unlabeled_examples_iter)
except StopIteration:
# if finished unlabeled data stream, restart it
unlabeled_examples_iter = unlabeled_data.batch_iter(
batch_size=args.batch_size, shuffle=True)
unlabeled_examples = next(unlabeled_examples_iter)
unlabeled_examples = [
e for e in unlabeled_examples
if len(e.tgt_actions) <= args.decode_max_time_step
]
unsup_encoder_loss, unsup_decoder_loss, meta_data = jae.get_unsupervised_loss(
unlabeled_examples, args.moves)
if bi_direction:
unsup_encoder_loss_back, unsup_decoder_loss_back, meta_data_back = jae.get_unsupervised_loss_backward(
unlabeled_examples, args.moves)
nan = False
if nn_utils.isnan(sup_loss.data):
print('Nan in sup_loss')
nan = True
if nn_utils.isnan(unsup_encoder_loss.data):
print('Nan in unsup_encoder_loss!', file=sys.stderr)
nan = True
if nn_utils.isnan(unsup_decoder_loss.data):
print('Nan in unsup_decoder_loss!', file=sys.stderr)
nan = True
if bi_direction:
if nn_utils.isnan(unsup_encoder_loss_back.data):
print('Nan in unsup_encoder_loss_back!', file=sys.stderr)
nan = True
if nn_utils.isnan(unsup_decoder_loss_back.data):
print('Nan in unsup_decoder_loss_back!', file=sys.stderr)
nan = True
if nan:
continue
if bi_direction:
report_unsup_encoder_loss += (
unsup_encoder_loss.sum().data[0] +
unsup_encoder_loss_back.sum().data[0])
report_unsup_decoder_loss += (
unsup_decoder_loss.sum().data[0] +
unsup_decoder_loss_back.sum().data[0])
else:
report_unsup_encoder_loss += unsup_encoder_loss.sum().data[0]
report_unsup_decoder_loss += unsup_decoder_loss.sum().data[0]
report_unsup_examples += unsup_encoder_loss.size(0)
if bi_direction:
unsup_loss = torch.mean(unsup_encoder_loss) + torch.mean(
unsup_decoder_loss) + torch.mean(
unsup_encoder_loss_back) + torch.mean(
unsup_decoder_loss_back)
else:
unsup_loss = torch.mean(unsup_encoder_loss) + torch.mean(
unsup_decoder_loss)
loss = sup_loss + args.unsup_loss_weight * unsup_loss
loss.backward()
# clip gradient
grad_norm = torch.nn.utils.clip_grad_norm(jae.parameters(),
args.clip_grad)
optimizer.step()
if train_iter % args.log_every == 0:
print(
'[Iter %d] supervised: encoder loss=%.5f, decoder loss=%.5f'
% (train_iter, report_encoder_loss / report_examples,
report_decoder_loss / report_examples),
file=sys.stderr)
print(
'[Iter %d] unsupervised: encoder loss=%.5f, decoder loss=%.5f, baseline loss=%.5f'
% (train_iter,
report_unsup_encoder_loss / report_unsup_examples,
report_unsup_decoder_loss / report_unsup_examples,
report_unsup_baseline_loss / report_unsup_examples),
file=sys.stderr)
samples = meta_data['samples']
for v in meta_data.values():
if isinstance(v, Variable): v.cpu()
for i, sample in enumerate(samples[:1]):
print('\t[%s] Source: %s' %
(sample.idx, ' '.join(sample.src_sent)),
file=sys.stderr)
print('\t[%s] Code: \n%s' % (sample.idx, sample.tgt_code),
file=sys.stderr)
ref_example = [
e for e in unlabeled_examples
if e.idx == int(sample.idx[:sample.idx.index('-')])
][0]
print('\t[%s] Gold Code: \n%s' %
(sample.idx, ref_example.tgt_code),
file=sys.stderr)
print(
'\t[%s] Log p(z|x): %f' %
(sample.idx, meta_data['encoding_scores'][i].data[0]),
file=sys.stderr)
print('\t[%s] Log p(x|z): %f' %
(sample.idx,
meta_data['reconstruction_scores'][i].data[0]),
file=sys.stderr)
print('\t[%s] Encoder Loss: %f' %
(sample.idx, unsup_encoder_loss[i].data[0]),
file=sys.stderr)
print('\t**************************', file=sys.stderr)
report_encoder_loss = report_decoder_loss = report_examples = 0.
report_unsup_encoder_loss = report_unsup_decoder_loss = report_unsup_baseline_loss = report_unsup_examples = 0.
print('[Epoch %d] epoch elapsed %ds' %
(epoch, time.time() - epoch_begin),
file=sys.stderr)
# perform validation
print('[Epoch %d] begin validation' % epoch, file=sys.stderr)
eval_start = time.time()
eval_results = evaluation.evaluate(dev_set.examples,
encoder,
args,
verbose=True)
encoder.train()
dev_acc = eval_results['accuracy']
print('[Epoch %d] code generation accuracy=%.5f took %ds' %
(epoch, dev_acc, time.time() - eval_start),
file=sys.stderr)
is_better = history_dev_scores == [] or dev_acc > max(
history_dev_scores)
history_dev_scores.append(dev_acc)
if is_better:
patience = 0
model_file = args.save_to + '.bin'
print('save currently the best model ..', file=sys.stderr)
print('save model to [%s]' % model_file, file=sys.stderr)
jae.save(model_file)
# also save the optimizers' state
torch.save(optimizer.state_dict(), args.save_to + '.optim.bin')
elif epoch == args.max_epoch:
print('reached max epoch, stop!', file=sys.stderr)
exit(0)
elif patience < args.patience:
patience += 1
print('hit patience %d' % patience, file=sys.stderr)
if patience == args.patience:
num_trial += 1
print('hit #%d trial' % num_trial, file=sys.stderr)
if num_trial == args.max_num_trial:
print('early stop!', file=sys.stderr)
exit(0)
# decay lr, and restore from previously best checkpoint
lr = optimizer.param_groups[0]['lr'] * args.lr_decay
print('load previously best model and decay learning rate to %f' %
lr,
file=sys.stderr)
# load best model's parameters
jae.load_parameters(args.save_to + '.bin')
if args.cuda: jae = jae.cuda()
# load optimizers
if args.reset_optimizer:
print('reset to a new infer_optimizer', file=sys.stderr)
optimizer = torch.optim.Adam(
[p for p in jae.parameters() if p.requires_grad], lr=lr)
else:
print('restore parameters of the optimizers', file=sys.stderr)
optimizer.load_state_dict(
torch.load(args.save_to + '.optim.bin'))
# set new lr
for param_group in optimizer.param_groups:
param_group['lr'] = lr
# reset patience
patience = 0