What's a free program that types audio in realtime?
If you mean generating subtitles, this is one of the few tasks where LLM AI does a great job, but you need some powerful hardware.
https://github.com/Purfview/whisper-standalone-win
>>1522956can i run this with cpu only no gpu?
>>1522958Haven't tried myself, but the repo mentions switching to GPU so probably
>>1522960I mean I haven't tried using CPU, the program has given me good results in the past
>>1522962thank you
>>1522960is not working or too slow, tried in my 17yo vaio a 10 sec flac file and after 10 minutes of 99% cpu load i closed the program
>>1523208On slower hardware you need to use a smaller model, e.g. "--model tiny.en", or just tiny if the audio isn't English.Some Vaio laptops actually have nvidia graphics cards, so you might try "--device cpu" to force CPU.Lastly, "--language English" speeds things up massively if you're not using a language specific modelThe tiny model is noticeably worse than the default medium, but it's 30x faster on CPU for me, while small is 15x faster.
>>1522632https://github.com/KoljaB/RealtimeSTT
>>1522632sherpa-onnx
