gabluz/ggml-whisper-large-quantized

How to run this ggml file?

WARNING: this can be slow and CPU intensive

Clone whisper.cpp repository:

git clone https://github.com/ggerganov/whisper.cpp

cd whisper.cpp

make

place this model called large_q5_0.bin inside whisper.cpp/models folder and run:

./main -m models/large_q5_0.bin yourfilename.wav

Command to transcribe to SRT subtitle files:

./main -m models/large_q5_0.bin yourfilename.wav --output-srt --print-progress

Command to transcribe to TRANSLATED (to English) SRT subtitle files:

./main -m models/large_q5_0.bin yourfilename.wav --output-srt --print-progress --translate

It can transcribe ONLY wav files!

Command line to convert mp4 (works for any video, just change the extension) to wav:

ffmpeg -i yourfilename.mp4 -vn -acodec pcm_s16le -ar 16000 -ac 2 yourfilename.wav

Command to convert all mp4 (works for any video, just change the extension) inside folder to wav:

find . -type f -iname ".mp4" -exec bash -c 'ffmpeg -i "$0" -vn -acodec pcm_s16le -ar 16000 -ac 2 "${0%.}.wav"' {} ;