model - 使用 pytorch 模型执行 SGD 的循环中的错误

Question

在用于执行随机梯度下降的函数中，我在所有批次和批次中的所有输入-输出对上都有这个循环：

train_points_loader = DataLoader(train_points, batch_size=64, shuffle=True, num_workers=1)

for batch, (X, y) in enumerate(train_points_loader):
  o.zero_grad() # setting gradient to zeros    
  pred = model(X) # get predictions through forward pass
  loss = l(pred, y) # compute the loss

  loss.backward() # backward propagation        
    
  o.step() # update the gradient to new gradients

我正在使用以下 pytorch 模型：

 model = nn.Sequential( 
 nn.Flatten(), 
 nn.Linear(input_size, hidden_layers[0]),
 nn.ReLU(),
 nn.Linear(hidden_layers[0], hidden_layers[1]),
 nn.ReLU(),
 nn.Linear(hidden_layers[1], output_size)
 )

当我运行函数内部的循环时，错误是：

  <ipython-input-24-a177ccf6125b> in train_loop(train_data_loader, 
  model, epochs, learning_rate, optimizer, loss)
  21     running_accuracy = 0.
  22     # loop over all batches and over all input-output pair within a 
  batch
  ---> 23     for batch, (X, y) in enumerate(train_data_loader):
  24 
  25         o.zero_grad() # setting gradient to zeros

  ValueError: too many values to unpack (expected 2) 

谢谢您的回答