I just want to raise this here. Also including the README.
A lot of parts are still TF specific. And many of those even don't mention that.
PyTorch relevant documentation is very minimal, lacking in many aspects.
Some PyTorch examples are missing.
The main README, and basic overview describes many not-so-relevant aspects, or even wrong statements when you consider the PyTorch backend (e.g. custom LSTM kernel is currently TF only, but also not so relevant nowadays, ...).
Feel free to directly update smaller changes. Or extend missing things.
For bigger restructuring, better let's discuss (briefly) before.
(cc @NeoLegends, not sure who else would (should) feel responsible for this...?)
I just want to raise this here. Also including the README.
A lot of parts are still TF specific. And many of those even don't mention that.
PyTorch relevant documentation is very minimal, lacking in many aspects.
Some PyTorch examples are missing.
The main README, and basic overview describes many not-so-relevant aspects, or even wrong statements when you consider the PyTorch backend (e.g. custom LSTM kernel is currently TF only, but also not so relevant nowadays, ...).
Feel free to directly update smaller changes. Or extend missing things.
For bigger restructuring, better let's discuss (briefly) before.
(cc @NeoLegends, not sure who else would (should) feel responsible for this...?)