documenting.txt

did not do normalize_audio_loudness as of now due to it feeling unnecessary and might lower the loudness pitch levels that may tamper with the model's efficiency
retained the punctuation in normalize_transcripts for similar reasons, ?, ! are important for guessing the tone

setting up espeak backend and configuring it for Phonemizer was hell, but it got solved by adding two lines.
The second line was found in the Phonemizer docs (bootphon)

for phoneme generation I am using hindi for bhojpuri's fallback since it is not supported out of the box by Phonemizer itself.

select the interpreter for the virtual environment before starting

encoding issues with putting transcript into excel files
i'll include how i solved them
instead of utf-8, i used utf-8-sig