I noticed that the current integration uses Faster Whisper for transcription. I would like to suggest replacing it with “Insanely Fast Whisper” for improved performance, especially in GPU-based environments. This version utilizes optimizations like Flash Attention 2, potentially leading to faster transcription times. Could this replacement be considered, or would it cause compatibility issues?
I noticed that the current integration uses Faster Whisper for transcription. I would like to suggest replacing it with “Insanely Fast Whisper” for improved performance, especially in GPU-based environments. This version utilizes optimizations like Flash Attention 2, potentially leading to faster transcription times. Could this replacement be considered, or would it cause compatibility issues?
https://github.com/Vaibhavs10/insanely-fast-whisper