def __init__(self):
self.path_train = config.path_train
self.path_test = config.path_test
self.path_dev = config.path_dev
custom_vocab = vocab.create_vocab()
custom_vocab.set_data(mode=config.all_mode)
self.vocab_list = custom_vocab.vocab_list
self.word_to_index = custom_vocab.word_to_index
self.index_to_word = custom_vocab.index_to_word
self.stop_words = stopwords.words('english')
self.labels = config.labels
self.label_to_index = {
self.labels[0]: 0,
self.labels[1]: 1,
self.labels[2]: 2
}
def __init__(self):
path = "./train_data.csv"
vocab = v.create_vocab(path_csv=path)
word_to_index = vocab.get_data()
self.dataset = c.Custom_dataset(word_to_index, path)
self.model = torch.load("./model.pth")
label = Variable(label.to(device))
sent = Variable(torch.stack(sent).to(device))
out = model(sent)
_, pred = torch.max(out.data, 1)
total += label.size(0) # batch size
correct += (pred == label).sum()
acc = 100 * (correct.cpu().numpy() / total)
return acc
if __name__ == "__main__":
path_csv = config.path_csv
# 데이터 처리
start = time.time()
vocab = v.create_vocab(path_csv=path_csv)
word_to_index = vocab.get_data()
print("time vocab load : ", time.time() - start)
start = time.time()
dataset = custom_dataset.Custom_dataset(word_to_index, path_csv=path_csv)
train_data = dataset.get_data()
print("데이터 준비 완료")
print("time data load : ", time.time() - start)
print(len(train_data))
train_loader = DataLoader(
train_data,
batch_size=config.batch,
shuffle=True,
label = Variable(label.to(device))
sent1 = Variable(torch.stack(sent1).to(device))
sent2 = Variable(torch.stack(sent2).to(device))
out = model(sent1, sent2)
_, pred = torch.max(out.data, 1)
total += label.size(0) # batch size
correct += (pred == label).sum()
acc = 100 * (correct.cpu().numpy() / total)
return acc
if __name__ == "__main__":
# 데이터 처리
start = time.time()
vocab = v.create_vocab(mode=config.vocab_mode)
vocab_list, word_to_index = vocab.get_data()
print("time vocab load : ", time.time() - start)
start = time.time()
glove = custom_glove()
embedding = glove.get_data(vocab_list)
print("time glove emb load : ", time.time() - start)
start = time.time()
dataset = custom_dataset.Custom_dataset(vocab_list, word_to_index)
train_data, test_data, dev_data = dataset.get_data()
print("time data load : ", time.time() - start)
train_loader = DataLoader(train_data,
batch_size=config.batch,