This repo intends to introduce a complete routine for deploying deep learning models to k210 board, using the MNIST digit recognition as an example.
Pytorch --> ONNX --> NNCASE --> KModel --> K210 SDK / MicroPython
You can find the tutorial here (简体中文), bon appetite!
This project is still building and improving, hoping to help more starters with deploying DL models on K210.
- Use Micropython to load KModel
- Use C SDK to load the KModel
- Add object detection pipeline
-
Gravati Open Dataset
- Website https://gas.graviti.com/
- MNIST Dataset https://gas.graviti.com/dataset/qiangzibro/MNIST
-
Datawhale
- Wechat official account: Datawhale
- Github https://github.com/Datawhalechina