Currently train.py is overloaded with many independent features. Understanding the code and what will be the result of a training call requires untangling the entire script. It's also an error prone UX. This is a first step at separating independent parts into their own scripts. |
||
---|---|---|
.. | ||
deepspeech_training |