Open SourceVoice & Media17k
faster-whisper
CTranslate2 reimplementation of Whisper at ~4x speed, lower memory.
faster-whisper is a reimplementation of OpenAI's Whisper inference using the CTranslate2 engine, delivering substantially faster transcription with lower memory use at the same accuracy. It supports quantization (e.g. int8) and efficient CPU and GPU execution, making it a common drop-in backend for Whisper-based tooling.
Repository
SYSTRAN/faster-whisperLanguage
PythonWhat you'd build with it
- Run Whisper transcription faster and cheaper in production speech-to-text services
- Deploy speech recognition on resource-constrained or CPU-only hardware via quantized models
- Serve as the transcription backend inside larger voice/agent pipelines
Tags
sttfastself-hosted