Deprecated code in dependencies

[mbarnig]

When training a model, typical warnings about deprecated code are shown below:

 > TRAINING (2021-12-02 10:54:53) 
/home/mbarnig/COQUI-TTS/TTS/TTS/tts/models/tacotron2.py:268: UserWarning: __floordiv__ is deprecated, and its behavior will change in a future version of pytorch. It currently rounds toward 0 (like the 'trunc' function NOT 'floor'). This results in incorrect rounding for negative values. To keep the current behavior, use torch.div(a, b, rounding_mode='trunc'), or for actual floor division, use torch.div(a, b, rounding_mode='floor').
  alignment_lengths = (
/home/mbarnig/COQUI-TTS/lib/python3.8/site-packages/torch/functional.py:445: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at  ../aten/src/ATen/native/TensorShape.cpp:2157.)
  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]

What should we do in such cases ?

1. ignore it
2. hope that the code is updated by the Coqui-TTS maintainers
3. modify the code by ourselves and submit a pull request

[mbarnig]

Referring to the answers given in the first Coqui-TTS online meeting, my understanding concerning deprecated code is the following :

As a simple user we can ignore these warnings.

If the behavior of an imported function into Coqui-TTS will change in an upcoming version of the related dependency, the maintainers of Coqui-TTS will take care of it.

A contributor who updates and tests himself the concerned Coqui-TTS Python script and asks for a pull request is always welcome