RuntimeError: Found dtype Long but expected Float

最新推荐文章于 2024-05-25 22:21:23 发布

唐僧爱吃唐僧肉

最新推荐文章于 2024-05-25 22:21:23 发布

阅读量8.2k

点赞数 7

分类专栏：报错问题解决文章标签： pytorch 深度学习神经网络

本文链接：https://blog.csdn.net/znevegiveup1/article/details/124778676

版权

报错问题解决专栏收录该内容

23 篇文章 1 订阅

订阅专栏

今天在跑模型的时候，发现了以下的报错

RuntimeError: Found dtype Long but expected Float

这里的报错发生在BCELoss的部分

loss_fn = torch.nn.BCELoss()
logit = logit.view(logit.size()[0]*logit.size()[1],-1)
batch_label = batch_label.view(batch_label.size()[0]*batch_label.size()[1],-1)
crossentropyloss = loss_fn(logit,batch_label)

这里的batch_label类型为torch.float，而crossentropyloss要求batch_label必须为torch.int类型
因此这段代码需要修改一下：

loss_fn = torch.nn.BCELoss()
logit = logit.view(logit.size()[0]*logit.size()[1],-1)
batch_label = batch_label.view(batch_label.size()[0]*batch_label.size()[1],-1)
batch_label = batch_label.to(torch.float)
crossentropyloss = loss_fn(logit,batch_label)