缩进错误。我完全不知所措

Question

我已经手动将整个事情间隔开。尽管如此，它还是行不通。第一行之后直接出现缩进错误。代码：def train_epoch(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples):""" docstring? """losses = []for d in data_loader:    input_ids = d['input_ids'].to(device)    targets = d['targets'].to(device)    outputs = model(input_ids = input_ids, labels = targets)    loss = loss_fn(outputs, targets)    losses.append( loss.item() )    loss.backward()    optimizer.step()    scheduler.step()    optimizer.zero_grad()return np.mean(losses)错误：def train_epoch(模型、data_loader、loss_fn、优化器、设备、调度器、n_examples): ...文件“”，第 2 行 ^ IndentationError：需要缩进块到底是怎么回事？我没有看到任何地方有问题。

Answer 1

def train_epoch(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples):

    """

    docstring?

    """

    losses = []

    for d in data_loader:

        input_ids = d['input_ids'].to(device)

        targets = d['targets'].to(device)

        outputs = model(input_ids = input_ids, labels = targets)

        loss = loss_fn(outputs, targets)

        losses.append( loss.item() )

        loss.backward()

        optimizer.step()

        scheduler.step()

        optimizer.zero_grad()

    return np.mean(losses)

请像这样格式化。您的代码不在您的train_epoch()方法的缩进块下。

Answer 2

我修改了它：for循环必须向右缩进才能使其函数代码运行或在函数内部，否则它不会被视为函数代码。此外，在这种情况下，损失列表必须位于 for 循环内部，而不是在其之前，也不能位于同一 for 循环级别。尝试一下然后告诉我。如果成功投票并回答:-)

def train_epoch(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples):

   

      

      for d in data_loader:

        losses = []

        input_ids = d['input_ids'].to(device)

        targets = d['targets'].to(device)

        outputs = model(input_ids = input_ids, labels = targets)

        loss = loss_fn(outputs, targets)

        losses.append( loss.item() )

        loss.backward()

        optimizer.step()

        scheduler.step()

        optimizer.zero_grad()

        return np.mean(losses)

Answer 3

这是因为第一行。函数体必须缩进。

def train_epoch(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples):

    """ docstring? """

    losses = []

    for d in data_loader:

        input_ids = d['input_ids'].to(device)

        targets = d['targets'].to(device)

        outputs = model(input_ids = input_ids, labels = targets)

        loss = loss_fn(outputs, targets)

        losses.append( loss.item() )

        loss.backward()

        optimizer.step()

        scheduler.step()

        optimizer.zero_grad()

    return np.mean(losses)