#Youtube video transcriptor based on OpenAI Whisper model. (Python Script)
45 messages · Page 1 of 1 (latest)
Can you provide a bit more detail on this? What is it doing? Is it grabbing a specific youtube link video and then providing the vocal audio in text?
After run the script you will enter any youtube video URL, pytube will turn the video into an mp3 file and Whisper will get the same file and transcribe it. Yes, it is only providing the audio text.
Example of eminem RAP GOD shorts
Before you have it run, you need to install first via Python:
pip install pytube
pip install openai-whisper
pip install ffmpeg
@bitter talon Tried the script. It's not working on my end.
I also suggest you add multilinear support, especially great for transcription of longer input audio.
Reference to init.py
@bitter talon Here's the issue I made on your new work. It's promising for enthusiasts!
@fading wharf Try python3 script.py
@fading wharf i'll fix it, thxx
okay in ordinary command prompt but not powershell
If you install whisper directly, it will not download all the whisper dependecies
downloaded and instslled it
Still having the error?
python3 would need me to install Python from the Windows Store.
but py would execute the script directly. I'm using Python 3.10
@fading wharf
Run this:
pip install git+https://github.com/openai/whisper.git pytube
where to run? because running it at the folder for Transcriptor it will return this error
I understand not everyone needs to install GitHub on their PC just for it to work.
@fading wharf Try this one: pip install -U openai-whisper
You are receiving errors because your PC don't have all the dependencies to run whisper. The error above is because you don't have git.
OpenAI whisper install doc: https://github.com/openai/whisper
...And it worked for me
Yes, as said earlier I believe not everyone has Git.
@fading wharf
We are committed to fostering a positive atmosphere and have implemented additional filters to block any inappropriate or disruptive language and topics.
Reminder: Refrain from bypassing the AutoMod filter as per server Rule #10.
Also YouTube URLs
They can be expanded so shortened youtube and youtube-music URLs are also allowed
Tried downloading a short, after downloading OpenAI whisper
Transcription could not start
@fading wharf Maybe you don't have all the dependecies. Try to install those tha are necessary to Whisper:
choco install ffmpeg - This one is for windows
pip install setuptools-rust
Even after installing the error persists
Choco?
No
Looks like good work has put on this. @bitter talon . May I know if it’s all ok to use. I also have a recent issue with ffmpeg . Does anyone has done this on windows. Pls let me know the right file to download for this. Looks like we need to give the path in environmental variable. @fading wharf any inputs on this, looks like u tested this
You have to install all the dependencies with Python first before using it. But it's broken on my end. ffmpeg, openai-whisper, pytube
To the effect, it didn't work at all on my system