liuyoude

The class_num is set as the number of IDs in train dataset, and if only use development dataset the code related to getting the training labels also needs to be...

> Has this problem been solved? file path should be set true

权重文件训练时模型的输出类别数量为41（6个machine type的总id数量），该报错是因为当前输出类别数量设置为1，输出类别数量参考该行代码：https://github.com/liuyoude/STgram-MFN/blob/d6c886559ed3d2f0ff3e35b289d391addcedbd7d/run.py#L40C51-L40C51

这种情况代码并不适用，需要有额外id数据进行分类，可以参考一下DCASE 2020技术报告中的代理异常方法看看

I also commit this problem, the result is terrible.

Can you get a good result?

> @wesbz @liuyoude One problem I found is that the loss of generator is significantly larger than the discriminator, for example: 6199809.67 vs 2.35. This might cause the non-convergence of...

> I think it may be caused by the weight set of generator loss