TimeseriesGenerator for many-to-many architectures

keras-team / keras-preprocessing

Utilities for working with image data, text data, and sequence data.

Other

1.02k stars 444 forks source link

TimeseriesGenerator for many-to-many architectures #287

Open kopytjuk opened 4 years ago

kopytjuk commented 4 years ago

What do you think about introducing support for "many-to-many" architectures (see here). In https://github.com/keras-team/keras/issues/11293 there is an issue about that functionality.

In the left side we have current implementation, on the right side the desired behaviour.

I would also love to contribute!

herzog-ch commented 4 years ago

Yes please, this feature would help a lot in training RNNs.

annaformaniuk commented 4 years ago

Good idea - for me this would also be a helpful utility, replacing a lot of preprocessing code.

nacheteam commented 4 years ago

I have been looking how to do this in a generator fashion, any examples of how to do this with the actual functions? I suppose I need to extend the Sequence class but I'm not pretty sure how to do it... Thanks in advance!