666-zhang
666-zhang
感谢回复,我输入的是如下的数据格式: 在model.fit的步骤会报错,在其中 `# 计算损失 logits, loss = self._get_train_loss(inputs, outputs, **kwargs)` 这一步时会显示输入数据的shape与期待不符: `ValueError: Using a target size (torch.Size([1, 100, 1])) that is different to the input size (torch.Size([1, 100])) is deprecated....
我将输入数据改为多条: 在`model.fit`阶段没有问题,在predict单条text的时候只返回一个空的集合, 请问predict的时候对于text的格式有什么要求吗?或者说怎样能得到和输入的label一样格式的输出?
收到谢谢
Hey sorry to bother you @Leavingseason , I'm an undergraduate student from Shanghai University of International Business and Economic School. I'm right now doing some research on recommendation system and...