5527c532288b90c6a52c3bab08a2832984d2d96c,classification/train_classifier.py,,eval_model,#Any#Any#Any#Any#,64

Before Change

    cnt = 0
    total_loss = 0.0
    for x, y in zip(valid_x, valid_y):
        x, y = Variable(x, volatile=True), Variable(y)
        output = model(x)
        loss = criterion(output, y)
        total_loss += loss.item()*x.size(1)
        pred = output.data.max(1)[1]

After Change

        return self.out(output)

def eval_model(niter, model, valid_x, valid_y):
    with torch.no_grad():
        model.eval()
        N = len(valid_x)
        criterion = nn.CrossEntropyLoss()
        correct = 0.0
        cnt = 0
        total_loss = 0.0
        for x, y in zip(valid_x, valid_y):
            output = model(x)
            loss = criterion(output, y)
            total_loss += loss.item()*x.size(1)
            pred = output.data.max(1)[1]
            correct += pred.eq(y).sum().item()
            cnt += y.numel()
        model.train()
    return 1.0-correct/cnt

def train_model(epoch, model, optimizer,
        train_x, train_y, valid_x, valid_y,

Instances