It is not a released tool, but only a snapshot of a private repository. I'm not adding documentation yet, however everything is self-explanatory.
To install pre-commit hooks:
pip install pre-commit
pre-commit install
To fine-tune ASR models:
python train.py --config=configs/base_config.py
A lot of useful information is saved:
- Loss/WER plots
- Model predictions
- Weight dynamics
- Training logs
The file configs/base_config.py
contains all details such as:
- The model checkpoint to use (Wav2vec2 or Whisper)
- Training and evaluation datasets
- Training hyperparameters
- Path to save results
See example run in example
folder.