A pure Rust implementation of Mistral's Voxtral Mini 4B speech recognition model using the Burn ML framework. The project features a Q4 GGUF quantized version (2.5 GB) that runs entirely client-side in browsers via WASM and WebGPU. It includes both native CLI and browser demo implementations, with custom solutions for WASM

5m read timeFrom github.com
Post cover image
Table of contents
Quick StartArchitectureBuildingTestingModel PreparationBenchmarksProject StructureLicense

Sort: