faster-whisper is a reimplementation of OpenAI’s Whisper automatic speech recognition model built around CTranslate2, an inference engine optimized for Transformer models.
Its primary aim is to provide Whisper transcription with lower processing times and more efficient memory usage than the original PyTorch implementation while retaining equivalent recognition behaviour. It can run on CPUs or NVIDIA GPUs and supports several computation types, including INT8 quantization for systems where memory consumption is particularly important.
The library is aimed at developers who want to incorporate Whisper into Python applications rather than providing a standalone desktop interface. It handles model acquisition, audio decoding, language detection and segmented transcription through a concise Python API.
This is free and open source software.
Key Features
- Runs OpenAI Whisper models through the optimized CTranslate2 inference engine.
- CPU inference for systems without a supported discrete graphics processor.
- NVIDIA GPU execution using CUDA, cuBLAS and cuDNN libraries.
- INT8 quantization on CPU to reduce memory requirements and accelerate suitable workloads.
- INT8 and mixed computation modes for supported GPU inference configurations.
- FP16 execution for efficient transcription on compatible NVIDIA graphics hardware.
- Batched inference pipeline for increasing throughput when processing suitable workloads.
- Compatible with Distil-Whisper checkpoints designed for more efficient Whisper inference.
- Automatic language detection with a probability value returned alongside transcription information.
- Segment-level timestamps identifying the beginning and end of recognised passages.
- Optional word-level timestamps for applications requiring more precise timing information.
- Integrated Silero voice activity detection for filtering portions of recordings without speech.
- Voice activity detection parameters can be adjusted to suit different silence characteristics.
- Uses PyAV for audio decoding, avoiding a requirement for a separately installed FFmpeg executable.
- Automatically downloads corresponding converted models when a model size is specified.
- Can load CTranslate2 Whisper models directly from a local directory.
- Conversion utilities support original Whisper and compatible Transformers model checkpoints.
- Works with fine-tuned Whisper models after conversion to the CTranslate2 model format.
- Python generator-based transcription interface allows segments to be consumed progressively.
- Supports beam-search configuration and other decoding parameters exposed by the transcription API.
Website: github.com/SYSTRAN/faster-whisper
Support:
Developer: SYSTRAN
License: MIT License
faster-whisper is written in Python. Learn Python with our recommended free books and free tutorials.
Related Software
| Speech Recognition Tools | |
|---|---|
| Whisper | Automatic speech recognition (system trained on 680,000 hours of data |
| Flashlight | Fast, flexible machine learning library written entirely in C++. |
| Coqui STT | Deep-learning toolkit for training and deploying speech-to-text models |
| Kaldi | C++ toolkit designed for speech recognition researchers. |
| SpeechBrain | All-in-one conversational AI toolkit based on PyTorch |
| Handy | Offline speech-to-text application |
| ESPnet | End-to-End speech processing toolkit |
| deepspeech.pytorch | Implementation of DeepSpeech2 using Baidu Warp-CTC. |
| Whispering | Transcription application with global speech-to-text functionality |
| Julius | Two-pass large vocabulary continuous speech recognition engine |
| CMUSphinx | Speech recognition system for mobile and server applications |
| Simon | Flexible speech recognition software |
| hyprwhspr | Native speech-to-text designed for Arch / Omarchy |
| ostt | Open Speech-to-Text |
| DeepSpeech | TensorFlow implementation of Baidu's DeepSpeech architecture. |
| OpenSeq2Seq | TensorFlow-based toolkit for sequence-to-sequence models |
| Eesen | End-to-End Speech Recognition |
Read our verdict in the software roundup.
 Explore our comprehensive directory of recommended free and open source software. Our carefully curated collection spans every major software category.This directory is part of our ongoing series of informative articles for Linux enthusiasts. It features hundreds of detailed reviews, along with open source alternatives to proprietary solutions from major corporations such as Google, Microsoft, Apple, Adobe, IBM, Cisco, Oracle, and Autodesk. You’ll also find interesting projects to try, hardware coverage, free programming books and tutorials, and much more. Discovered a useful open source Linux program that we haven’t covered yet? Let us know by completing this form. |
Please read our Comment Policy before commenting.