import os
 
os.environ["CUDA_VISIBLE_DEVICES"] = "0,4"
ids = [0,1]

class CNN(nn.Module):
  def __init__(self):
    super(CNN,self).__init__()
    self.conv1 = nn.Sequential(
    ......
    )
    
    ......
    
    self.out = nn.Linear(Liner_input,2)
 
  ......
    
  def forward(self,x):
    x = self.conv1(x)
    ......
    output = self.out(x)
    return output,x
  
cnn = CNN()
 
# 更改，.cuda()表示将本存储到CPU的网络及其参数存储到GPU！
cnn.cuda()

for epoch in range(EPOCH):
  epoch_loss = 0.
  for i, data in enumerate(train_loader2):
    image = data['image'] # data是字典，我们需要改的是其中的image
 
    #############更改！！！##################
    image = Variable(image).float().cuda()
    ############################################
 
    label = inputs['label']
    #############更改！！！##################
    label = Variable(label).type(torch.LongTensor).cuda()
    ############################################
    label = label.resize(BATCH_SIZE)
    output = cnn(image)[0]
    loss = loss_func(output, label)  # cross entropy loss
    optimizer.zero_grad()      # clear gradients for this training step
    loss.backward()         # backpropagation, compute gradients
    optimizer.step() 
    ... ...

  ... ...
    #################################################
    pred_y = torch.max(test_train_output, 1)[1].data.cpu().numpy()
    
    accuracy = float((pred_y == label.cpu().numpy()).astype(int).sum()) / float(len(label.cpu().numpy()))

class VGG(nn.Module):
 
  def __init__(self, features, num_classes=2, init_weights=True):
    super(VGG, self).__init__()
... ...
 
  def forward(self, x):
    #x = self.features(x)
    #################Multi GPUS#############################
    x = nn.parallel.data_parallel(self.features,x,ids)
    x = x.view(x.size(0), -1)
    # x = self.classifier(x)
    x = nn.parallel.data_parallel(self.classifier,x,ids)
    return x
... ...