ggerganov / whisper.cpp

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

Add support for decoding using FFmpeg in the `main` example #2432

Open shih-liang opened 1 month ago

easyfab commented 1 month ago

Could be interesting to have this in main for those you don't want to install ffmpeg. But you can also simply use directlty ffmpeg to pipe into main. e.g. : ffmpeg -i file -ar 16000 -ac 2 -c:a pcm_s16le -f wav - | main -m model.bin -f -