错误1:

Expected object of backend CPU but got backend CUDA for argument #2 'weight'   

#之前:
    model = model.cuda()
#之后:
    model = torch.nn.DataParallel(model).cuda()

 

 

错误2:

IndexError: Dimension out of range (expected to be in range of [-1, 0], but got 1)

 #之前:
# 前向传播计算输出
            outputs = model(input)
            label = label.squeeze()
            outputs.detach_()
            # 计算准确率,记录损失
            loss = criterion(outputs[i], label)
#之后:
             loss = criterion(outputs, label)            

 

#之前label有一个维度问题 
label.size()=tensor([16,1])

label = label.squeeze()

label.size()=tensor([16])

 

posted on 2020-07-08 16:18  Manuel  阅读(146)  评论(0编辑  收藏  举报