Some makers sure know how to make a Raspberry Pi project you can really savor like developer Simon who’s created a Pi-powered AI meat cooking monitor known as the LLM Steak Classifier AIOT Bot. This project uses ChatGPT in conjunction with physical sensors to help you make sure the meat you’re cooking is completely done and ready to eat.
Simon explains that the project idea began when he became aware of the release of the GPT-4-Vision model. Wanting to explore this new technology, Simon decided to create a project that would use ChatGPT to test whether or not meat is thoroughly cooked using images of the meat, sensor data, and text prompts about the setup.
There are two Raspberry Pis used in this project—a Raspberry Pi 4 and a Raspberry Pi Pico compatible board. The Pi 4 is used to capture images of the meat and parse information to ChatGPT. It also hosts a Streamlit web server that users can use to access a custom interface with information about the meat's cooking progress and feedback from ChatGPT. This web server can be accessed by any device with a web browser. In the example, Simon is using a smartphone.
The Raspberry Pi 4 is connected to a W5100S-EVB-PICO module. This RP2040-based board is used to monitor the temperature of the meat using a DS18B20 temperature sensor. This data is used in tandem with images of the meat while it cooks to help determine whether or not it’s done. Temperature data is also displayed on the streamlit web server.
For those curious about how this Raspberry Pi project works, you’ll be happy to know Simon was kind enough to share the source code. You can explore the code behind this project over at the official project page at Wiznet. There you’ll also find a detailed breakdown of how the project is assembled and plenty of other details you can use to recreate it from scratch.
