Deep Learning with PyTorch

Training the model

Training the model is very similar to what we saw for building image classifiers, so we will be using the same functions. We pass batches of data through the model, calculate the outputs and losses, and then optimize the model weights, which includes the embedding weights. The following code does this:

def fit(epoch,model,data_loader,phase='training',volatile=False):    if phase == 'training':        model.train()    if phase == 'validation':        model.eval()        volatile=True    running_loss = 0.0    running_correct = 0    for batch_idx , batch in enumerate(data_loader):        text , target = batch.text , batch.label        if is_cuda:            text,target = text.cuda(),target.cuda()                if phase == 'training':            optimizer.zero_grad()        output = model(text) loss = F.nll_loss(output,target) ...