Details

We will be demonstrating the inference acceleration of AI models on Raspberry Pi using our AI inference engine, Optimium.

Raspberry Pi has been used for various AI applications such as object detection for delivery robots and LLMs including ChatGPT. Our team has conducted a variety of AI optimization projects to deploy computer vision models on Raspberry Pi. For example, we had a client who encountered a latency problem while deploying MediaPipe pose landmark model on Raspberry Pi 5. Additionally, we successfully accelerated inference speed of the client’s model on Raspberry Pi using our own inference engine, Optimium.

Optimium is a next-generation inference optimization engine. It accelerates AI model inference on target hardware while maintaining accuracy and facilitates convenient AI model deployment across various hardware platforms using a unified tool. Most importantly, it is highly optimized for Arm CPUs, which Raspberry Pi is utilizing.

In the webinar, we will be demonstrating the optimization of an AI model for Raspberry Pi. We are going to run the AI model optimized by Optimium on Raspberry Pi and compare the result with those from models optimized using alternative inference engines.

Visit organiser website