github.com
WhisperS2T is an optimized speech-to-text pipeline for the Whisper model that supports multiple inference engines. The repository provides release notes, quickstart guides, future roadmaps, and a benchmark and technical report.