AI / ML

tiny-audio, nanoGPT for speech-to-text

tiny-audio is a tiny (~1.5KB) audio library for Rust, which provides a simple API for playing and recording audio. It has been used in conjunction with nanoGPT, a lightweight version of the popular GPT model, to create a speech-to-text system. nanoGPT is a 70MB model that can be used for large-scale language tasks, and tiny-audio allows it to interact with audio inputs. The combination of the two allows for real-time speech recognition and can be used in applications such as voice assistants, chatbots, and more.

Read the full article at github.com →