What does faster mean and what advantage this have over original whisper?
#1
by
supercharge19
- opened
Does it output more tokens/words in same amount of time than original whisper? Does it require less amount of RAM/VRAM and compute (and how much)?
Hi. This is the large-v3 conversion model used for faster-whisper engine. The faster-whisper engine supports speech-to-text conversion up to 4x faster than openai/whisper with the same accuracy while using less memory.
For more information, you can refer to the benchmark here: https://github.com/SYSTRAN/faster-whisper?tab=readme-ov-file#benchmark