def test(config):
device = torch.device(config.device)
# load data
data_dir = os.path.join(config.model, config.dataset)
test_fname = os.path.join(data_dir, config.test_data)
test_data = get_loader(test_fname, config.batch)
wordemb = np.loadtxt(os.path.join(data_dir, config.wordmat_file))
targetemb = np.loadtxt(os.path.join(data_dir, config.targetmat_file))
# init model
model = ABSA_Atae_Lstm(dim_word=config.dim_word,
dim_hidden=config.dim_hidden,
num_classification=config.num_class,
maxlen=config.sent_limit,
wordemb=wordemb,
targetemb=targetemb,
device=device)
# load model
model_save_dir = os.path.join(config.model_save, config.dataset,
config.model)
result_dir = os.path.join(config.result_save, config.dataset, config.model,
'test')
if not os.path.exists(result_dir):
os.makedirs(result_dir)
save_fout = open(os.path.join(result_dir, 'best.txt'),
'w',
encoding='utf-8')
model.load_state_dict(torch.load(os.path.join(model_save_dir, 'best.pth')))
model = model.to(device)
model.eval()
# init loss
logit_list = []
rating_list = []
for batch_data in tqdm(test_data):
sent_ids, aspect_id, polarity, lens = batch_data
sent_ids, aspect_id, polarity, lens = sent_ids.to(device), aspect_id.to(device), \
polarity.to(device), lens.to(device)
logit = model(sent_ids, aspect_id, lens)
save(sent_ids.tolist(), lens.tolist(), aspect_id.tolist(),
polarity.tolist(), logit.tolist(), save_fout, config)
logit_list.append(logit.cpu().data.numpy())
rating_list.append(polarity.cpu().data.numpy())
test_acc, test_precision, test_recall, test_f1 = get_score(
np.concatenate(logit_list, 0), np.concatenate(rating_list, 0))
print(
' test_acc=%.4f, test_precision=%.4f, test_recall=%.4f, test_f1=%.4f' %
(test_acc, test_precision, test_recall, test_f1))
def train(config):
device = torch.device(config.device)
# random seed
if torch.cuda.is_available():
torch.cuda.manual_seed_all(config.seed)
# load data
data_dir = os.path.join(config.model, config.dataset)
train_fname = os.path.join(data_dir, config.train_data)
test_fname = os.path.join(data_dir, config.test_data)
train_data = get_loader(train_fname, config.batch)
test_data = get_loader(test_fname, config.batch)
wordemb = np.loadtxt(os.path.join(data_dir, config.wordmat_file))
targetemb = np.loadtxt(os.path.join(data_dir, config.targetmat_file))
# init model
model = ABSA_Atae_Lstm(dim_word=config.dim_word, dim_hidden=config.dim_hidden, num_classification=config.num_class,
maxlen=config.sent_limit, wordemb=wordemb, targetemb=targetemb, device=device)
model = model.to(device)
# init loss
cross_entropy = nn.CrossEntropyLoss()
# train
# summary writer
writer = SummaryWriter('logs/%s/%s/%s' % (config.dataset, config.model, config.timestr))
model_save_dir = os.path.join(config.model_save, config.dataset, config.model)
if not os.path.exists(model_save_dir):
os.makedirs(model_save_dir)
optim = torch.optim.Adagrad(model.parameters(), lr=config.lr)
best_acc = 0.0
for epoch in tqdm(range(config.max_epoch)):
# train
model.train()
for batch_data in tqdm(train_data):
model.zero_grad()
sent_ids, aspect_id, polarity, lens = batch_data
sent_ids, aspect_id, polarity, lens = sent_ids.to(device), aspect_id.to(device), \
polarity.to(device), lens.to(device)
logit = model(sent_ids, aspect_id, lens)
loss = cross_entropy(logit, polarity)
loss.backward()
optim.step()
# eval
model.eval()
# eval on train
logit_list = []
rating_list = []
for batch_data in tqdm(train_data):
sent_ids, aspect_id, polarity, lens = batch_data
sent_ids, aspect_id, polarity, lens = sent_ids.to(device), aspect_id.to(device), \
polarity.to(device), lens.to(device)
logit = model(sent_ids, aspect_id, lens)
# loss = cross_entropy(logit, polarity)
logit_list.append(logit.cpu().data.numpy())
rating_list.append(polarity.cpu().data.numpy())
train_acc, train_precision, train_recall, train_f1 = get_score(np.concatenate(logit_list, 0),
np.concatenate(rating_list, 0))
# writer.add_scalar('train_loss', train_loss, epoch)
writer.add_scalar('train_acc', train_acc, epoch)
writer.add_scalar('train_precision', train_precision, epoch)
writer.add_scalar('train_recall', train_recall, epoch)
writer.add_scalar('train_f1', train_f1, epoch)
# eval on test
logit_list = []
rating_list = []
for batch_data in tqdm(test_data):
sent_ids, aspect_id, polarity, lens = batch_data
sent_ids, aspect_id, polarity, lens = sent_ids.to(device), aspect_id.to(device), \
polarity.to(device), lens.to(device)
logit = model(sent_ids, aspect_id, lens)
# loss = cross_entropy(logit, polarity)
logit_list.append(logit.cpu().data.numpy())
rating_list.append(polarity.cpu().data.numpy())
test_acc, test_precision, test_recall, test_f1 = get_score(np.concatenate(logit_list, 0),
np.concatenate(rating_list, 0))
# writer.add_scalar('test_loss', test_loss, epoch)
writer.add_scalar('test_acc', test_acc, epoch)
writer.add_scalar('test_precision', test_precision, epoch)
writer.add_scalar('test_recall', test_recall, epoch)
writer.add_scalar('test_f1', test_f1, epoch)
print('epoch %2d : '
' train_acc=%.4f, train_precision=%.4f, train_recall=%.4f,train_f1=%.4f,'
' test_acc=%.4f, test_precision=%.4f, test_recall=%.4f, test_f1=%.4f' %
(epoch, train_acc, train_precision, train_recall, train_f1,
test_acc, test_precision, test_recall, test_f1))
# show parameters
for name, param in model.named_parameters():
writer.add_histogram(name, param, epoch, bins='doane')
# save model
torch.save(model.state_dict(), os.path.join(model_save_dir, '{}.pth'.format(epoch)))
if test_acc > best_acc:
torch.save(model.state_dict(), os.path.join(model_save_dir, 'best.pth'))
best_acc = test_acc