requirements

'onnx>=1.12.0,<2.0.0', 'onnxslim>=0.1.71', 'onnxruntime'

dataset used (for v2.0 model)

https://universe.roboflow.com/xs-qmjxi/grape-btpnv/dataset/1 epoch: 2, batch_size=16

train

download https://universe.roboflow.com/yogeshs-workspace-shhzv/grape-disease-detection-tlnao/dataset/1 copy data.yaml, modify the path for image to make it fits your file structure

python train.py

TODO

depends on time, not sure if we have time to code custom inference code to make it work with NPU