How to optimize custom HEF models

When using Hailo-8l custom model inference, there is a lot of noise. I would like to ask how to optimize it.


When I use the onnx model for detection, it is normal. The above problem occurs during detection when onnx is converted to HEF model.

May I ask everyone how to optimize HEF.

First I would recommend to work trough the tutorials in the Hailo AI Software Suite Docker. You can start them by running the following command:

hailo tutorial

This will start a Jupyter notebook server with notebooks for each step of the conversion process including Model Optimization and Layer Noise Analysis.

The Hailo Dataflow Compiler User Guide also contains useful information about the Model Optimization in chapter 5.3.