train_loader = DataLoader(d_train, batch_size=batch_size, shuffle=True)
val_loader = DataLoader(d_val, batch_size=batch_size, shuffle=False)
### Train model
print('Train LSTM ... ')
model = LSTMClassifier(embedding_dim, hidden_dim, num_layers, batch_size)
model.init_weights()
model.embedding.weight = torch.nn.Parameter(torch.Tensor(word2vec.wv.syn0))
model.embedding.weight.requires_grad = False
model.cuda()
print(model)
criterion = nn.NLLLoss()
optimizer = optim.Adam(
filter(lambda p: p.requires_grad, model.parameters()))
for epoch in range(4):
train(epoch, model, criterion, optimizer, train_loader)
validate(epoch, model, val_loader)
print('Done !')
### Save model
print('Save model ... ', end='')
torch.save(model.state_dict(), '_model.pt')
print('Done !')
outputs = model(words)
# print("output_shape")
# print(outputs.shape)
# if y would be one-hot, we must apply
# labels = torch.max(labels, 1)[1]
loss = F.cross_entropy(outputs, labels)
# Backward and optimize
optimizer.zero_grad()
loss.backward()
optimizer.step()
if (epoch + 1) % 100 == 0:
print(f'Epoch [{epoch+1}/{num_epochs}], Loss: {loss.item():.4f}')
print(f'final loss: {loss.item():.4f}')
data = {
"model_state": model.state_dict(),
"input_size": input_size,
"hidden_size": hidden_size,
"output_size": output_size,
"all_words": all_words,
"tags": tags
}
FILE = "lstmmodel.pth"
torch.save(data, FILE)
print(f'training complete. file saved to {FILE}')
