This is dirty version of extracting transcription from the Youtube video.
380
This is work-in-progress project. It is basically wrapper arround yt-dlp and openai-whisper. This project currently involved undocumented manual steps. It can be run both on Windows and Linux with or without GPU. Released configuration assumes GPU.
It can be run on bare metal via uv or in docker container both via docker and docker compose.
Not all functionality is docerized, there are also external scripts (for bash and for cmd) that relies on bare metal installations.
The project will be completely rewritten, but this task has low priority, because it does the job.
See https://github.com/alex-ber/AlexBerDirtyWinYt for source code.
Content type
Image
Digest
sha256:1a8adab36…
Size
8.4 GB
Last updated
about 1 month ago
docker pull alexberkovich/neural-extractor-node