issue about ValueError

[panxiaoyuandbishe]

Hello Dr.Mousavi,
I have two problems. Recently, when I use EQTransformer to process the data I downloaded, I always encounter the same error, ValueError(not enough values to unpack (expected 3, got 0)), this error appears in the model.predict_generator function of mseed-predictior(line315). The format of the mseed file is correct, and the input values ​​before the error appear also exist. I also tried changing the time length of the data, but always the same error.
Another problem is how the attention distribution map that appears in the article is drawn. I'm new to the field of deep learning, and I'm very unfamiliar with these visualizations. The information provided on the Internet has not been able to provide effective help.
Can you give some suggestions?

[smousavi05]

@panxiaoyuandbishe this is a version issue. I need to know how you have installed EqT and on which type of machine, mac or windows?
The second question is a bit more complicated. You need to extract the attention weights from the model and then do some sort of post-processing/stacking to get the attention map.

[panxiaoyuandbishe]

Thanks for your reply. I have Obspy installed on my machine so I get EQTransformer through PyPI. And the machine I use is Windows.

[smousavi05]

Try installing EqT directly from Github repo. Other versions won't work on Windows. @panxiaoyuandbishe