It's a great idea to give people a chance to reflect. Really like this! Technically, I"m just wondering, the app says 22 languages transcribed, offline, and it's ~480Mb in size. I thought only the large model (> 1Gb) had more than English (and the models don't compress). And when you say transcribed do you mean transcribed to English?
BTW - to modify an existing app, can you just fork the repo and open a new Xcode on that fork? I don't know iOS but I'm getting into MacOS dev and I like this idea of modifying an existing app, as I've already got so much boiletplate I think I can reuse.
For each size there is a ".en" and a multilingual one. I'm using the multilingual one -> you talk in German and receive a German transcript. On page 23 of the paper you can see the WER's of each language (https://cdn.openai.com/papers/whisper.pdf) I limited my app to the languages that had a max of around 20% WER.
Sure if the repo has a xcodeproj you can just open it in XCode, change the signing and improve on it. (Just make sure to always respect the licenses) If you want to play with whisper.cpp you can use this SwiftUI demo: https://github.com/ggerganov/whisper.cpp/tree/master/example...
Wow, thanks for the info and paper link! You are awesome. Also the assurance on how to fork projects. :) Good luck with your app! I'm also launching my Whisper related voice memo transcription app soon: https://apps.apple.com/app/wisprnote/id1671480366
BTW - to modify an existing app, can you just fork the repo and open a new Xcode on that fork? I don't know iOS but I'm getting into MacOS dev and I like this idea of modifying an existing app, as I've already got so much boiletplate I think I can reuse.