ibayer

Jep, feels like a memory leak. Not something that a higher memory limit can fix. But I didn't check the code and know very little about php... Anyway, thanks for...

Related https://github.com/intake/intake/issues/342 since a pytorch dataloader would require a file path and handle the loading from disk.

@martindurant Thanks, that's exactly the answer I was looking for. I'll keep this issue open in case we run into any roadblocks and report back with our experience later.

@danielskatz We have just been informed that an ipyannotator introduction talk, that @itepifanio have submitted to Python Nordeste, has been accepted. It would be great if we could incorporate feedback...

> I'm sorry, I don't understand what "if we could in incorporate feedback from this review" means. Ah, sorry I didn't mean to be cryptic. All I wanted to express...

Thanks for the feedback! > The significance of contributions by the first author to the source code base cannot be verified. We have been following a Private Development, Public Releases...

@csadorf https://github.com/palaimon/ipyannotator/pull/34 has been merged now. Does this now complete your last open checkbox? Please let us know if you have further requests before we get your final review approval.

The model parameter in fastFM can be inspected after each iteration without much overhead. This allows you to create your own logging output in python. This approach is much more...

Hi the 'n_more_iter` option allows to train the model over junks of data see http://ibayer.github.io/fastFM/guide.html#learning-curves, you just have to swap the training set at every iteration of the loop. fm.fit(X_train,...

@parklize your right warm start is currently not implemented for SGD (and BPR). I'm not sure, but it might be possible to add warm start support without changing the C...