Open PgvCMCz518 opened 1 month ago
Hi @PgvCMCz518, Thanks for bringing this up. Compared to insanely-fast-whisper, it seems like it uses quantization techs called HQQ and BNB.
I don't know how it's better than other implementations in terms of performance, But I'll run some benchmarks and would integrate it after confirming it's better at something when I have time.
https://github.com/kadirnar/whisper-plus seems great, thanks.