03 Edge AI
Waste Classification on a Raspberry Pi 5
2024
Lede
Vision classifier I squeezed onto a Pi 5 so it could run offline. Real-world stuff ended up mattering more than test accuracy.
Detail
Set out to build a small camera-based waste sorter that could run without a network connection. The model isn't the interesting part — the constraints are. Throttling under thermal load, INT8 quantization that broke parts of the network on first try, the slightly different image stats from a cheap CSI camera. The takeaway: most of the work in a deployment-first ML project is not actually the model.
Stack
- ONNX
- Raspberry Pi 5
- OpenCV
- INT8 quantization