whisper.cpp
LibraryOffline speech-to-text
Use it when you need offline speech-to-text in C/C++ on CPU, GPU or Apple Silicon
A C/C++ port of OpenAI's Whisper speech recognition model for fast, dependency-free transcription.
Added to the shelf
From the makersREADME, read 15 Oct
whisper.cpp is a high-performance C/C++ inference port of OpenAI's Whisper automatic speech recognition model. It has no dependencies, runs on CPU and on GPUs or NPUs including Metal, CUDA, Vulkan, Core ML and OpenVINO, and exposes a C-style API. It is built to run offline on many platforms, from Apple Silicon and Android to WebAssembly and Raspberry Pi.
Commands
git clone https://github.com/ggml-org/whisper.cpp.gitClone the repository
sh ./models/download-ggml-model.sh base.enDownload a ggml Whisper model
cmake -B buildConfigure the build
cmake --build build -j --config ReleaseBuild the project
./build/bin/whisper-cli -f samples/jfk.wavTranscribe an audio file
./build/bin/quantize models/ggml-base.en.bin models/ggml-base.en-q5_0.bin q5_0
How creators use itNone yet
No creator takes yet. This one came straight from its repo.