Celebrate with AI: Chinese New Year Tips from Mistral and LLaVA on Raspberry Pi
Introduction
Welcome to this article where the traditional festival meets cutting-edge technologies. As we approach the Chinese Lunar New Year, it is the perfect moment to enjoy the festival period with some cultural advice from our AI assistants, either with a Large language model from Mistral AI or a multi-modal model like LLaVA.
The specialty of this article is that we are going to deploy the mentioned models on a low-cost small-edge device – Raspberry Pi, making the advanced AI technology accessible even in your kitchen, as many other domestic appliances.
As we step into the year of the Dragon, it's an exciting moment to take advantage of a tiny AI model on a commodity device and enjoy the festive atmosphere. Don't worry if you have never used a Raspberry Pi or generative AI before; in this article, I will walk you through every single step of this project, from zero to hero.
Let's get started!
Hardware prerequisite
As you've understood, this article will require some basic Hardware. Here's what you'll need:
- Raspberry Pi 5–8 GB: I purchased the Raspberry Pi separately for 129 euros. Alternatively, you can buy a starter kit which may include all the components listed below.
- Micro Memory Card: The larger and faster the SD card, the better. I opted for the "SanDisk 128 GB MicroSDXC + SD Adapter with A2 App Performance Up to 190 MB/s, Class 10, U3, V30". This micro SD card cost me 27 euros.
- Micro HDMI Converter: Needed for the display. I bought the most affordable option for about 4 euros.
- Fan (Optional): To cool the CPU. Initially, I didn't purchase a fan, but as my CPU became excessively hot and my AI models began to run more slowly, I realized a cooling fan might improve performance. I just acquired one for 20 euros this morning, and you'll get to see my lovely fan in my next article.
