Closed chadcwilliams closed 4 months ago
Did you want to give a sequence of 82 data points to the GAN when it was constructed for and trained with sequences of length 100 ? In that case, this error would make sense
As far as I can remember, that error did not show up when training on other datasets - or were they sampled down to 100 data points as well?
This issue was created a long time ago, so it could just be I screwed up (as I wasn't familiar with the code back then). I kept it up here just to test it and make sure it's not a problem anymore.
aaaah okay I understand
This is no longer an issue.
I encountered an issue where inputting data into EEG-GAN training would crash unless the time-series were exactly 100 datapoints.