Closed OfflinePuffin closed 1 month ago
Even if I could generate lyrics, I still don't like the idea of implementing them. You can see my technical and kinda-ranty comment here for a really in-depth explanation on why I'm not comfortable with lyrics in any form.
Not only that, but pumping copyrighted music into a public API would not only break my "No internet connectivity" promise but also raise significant legal issues for the app.
Sorry.
I thought these models can be ran offline on device? But I do see your point. Thankyou for talking the time to respond
If they are, I imagine they are still quite intensive to run over time, especially regarding something like battery life. @OfflinePuffin
Probably, I thought it would work like the Manual full sync button, All songs would be processed with a manual sync, then a lyrics file would be generated for each file "both files will be named the same", And saved in a lyrics dir. Some logic could be added so that If it detects a lyric file that it's already been created based off of a file name for the song then it will skip it.
If there are new songs added it can be done on the fly... I know this issue is closed, I apologize
@OxygenCobalt
Description
Would the technique used by Alex Kostarevas over at FUTO for his live caption software be applicable to provide Song lyrics,
Using thr OpenAi whisper model
https://www.youtube.com/watch?v=gkCofPjEwOg
Problem solved
No response
Other implementations
Use a Lyric file
Benefit
To have generated lyrics for songs
Duplicates