While using Tacotron 2 I found myself spending a long time to set up the transcript file. So, I made a simple python script to download transcript & add those extra markers in every line.
How to use:
- Open command prompt by typing cmd in windows search
- type: cd Download_location
- type: python Tacotron-2-auto-Transcript.py
- use 1 to download transcript
- use 2 to reformat file.
python packages to install:
- youtube-transcript-api (pip install youtube-transcript-api)
What it does:
- It can download transcript from provided youtube video id & save it into a txt file.
- It adds /wav[number].wavs| {some txt} to every line.
- By default it downloads 'DY0ekRZKtm4' video transcript if you don't give an address.
What it doesn't do:
- If transcription lines aren't organized or not separated by full sentences it doesn't separate it.
- Doesn't add missing punctuationmarks. I tried to do it but it requires machine learning & I'm fairly new to coding. Infact this is my first original code.
What you can help with:
- You can help with automatic punctuation marks.
- Create a GUI & make it user friendly.