ggerganov / whisper.cpp

Port of OpenAI's Whisper model in C/C++
MIT License
35.38k stars 3.61k forks source link

Performance Issues with devices with low processing power #2482

Open tryWabbit opened 2 weeks ago

tryWabbit commented 2 weeks ago

Thank you so much for the incredible library.

I'm trying the SwiftUI Example, and it works perfectly fine. However, for long audios, let's say 60 minutes, it takes too long to process in devices like iPhone 11. It works fine with my iPhone 14 and later versions. It takes 25-35 minutes to complete in iPhone 11 whereas it takes 5-10 minutes in iPhone 14.

Can we make any optimizations to make the transcription process faster in older devices?

Again, Thanks in advance.