Nhat Pham

> Make sure that we provide the conversion back too. @conradoverta what do you mean by conversion back 🤔

> > > Make sure that we provide the conversion back too. > > > > > > @conradoverta what do you mean by conversion back 🤔 > > How...

@convoliution No worries, I was going to do it anyway. The conflict resolution might take awhile tho, since I don't rmb the details of the PR.

Conflicts resolved. That took less time than I thought.

> Hey @nhatsmrt have you gotten a chance to test this PR out? @convoliution Yes, but there were a few deployment bugs polluting the view. One is the comment you...

Waiting for documentation to be hosted.

This is not mentioned in the paper, but I think we can get around the problem of variable input size by combining this with an adaptive pooling layer right before...

You can create a Keras Sequence, in which you can apply the embedder on the input sequences, and use model.fit_on_generator/predict_on_generator