Transcribe audio from Youtube videos using Whisper.jl, a Julia package for Automatic Speech Recognition (ASR) that leverages OpenAI's Whisper model.
Requeriments:
yt-dlp
for downloading and extracting the audio.ffmpeg
for transcoding the audio.
Usage:
- First, download the audio of the video with the "Download" button.
- Next, transcribe the audio with the "Transcribe" button. This can take a while.
- To use a different Whisper model, edit the call to
transcribe
inwhisper.jl
Clone the repository and install the dependencies:
First cd
into the project directory then run:
$> julia --project -e 'using Pkg; Pkg.instantiate()'
Then run the app
$> julia --project
julia> using GenieFramework
julia> Genie.loadapp() # load app
julia> up() # start server
Open your browser and navigate to http://localhost:8000/
The Docker image for this app is available available at pjgim/whisper
and can be run with
docker run -p 8000:8000 pjgim/whisper
You can also build your own with the included Dockerfile.