The code is structured as two pipelines of scripts. The following diagrams capture the dependency structure of the scripts (the following script depends on the output of the previous script):
Install the listed dependencies for each of these modules -- following the instructions on each of their pages.
Link: https://github.com/huggingface/evaluate Library Version: 0.4.0 Python Version: 3.8
Link: https://github.com/pytorch/torcheval Library Version: 0.0.7 Python Version: 3.8
Link: https://github.com/m-bain/whisperX Version: 3.1.1
Link: https://platform.openai.com/docs/introduction Version: 1.9.0
Link: https://github.com/pariajm/english-fisher-annotations
We had to modify this code, so we provide the code here as a subdirectory.
Link: https://podcastsdataset.byspotify.com/
This dataset is maintained by Spotify, and access to the dataset is determined by Spotify.
- Pandas (Link: https://pandas.pydata.org/)
- tqdm (Link: https://github.com/tqdm/tqdm)