Whisper
Open-source speech recognition model with multilingual support
About
Whisper is an open-source automatic speech recognition model trained via large-scale weak supervision. It supports transcription and translation across over 100 languages and can be self-hosted under an MIT license.