Hailo-8 model conversion - CLI

Hi everyone,

I hope you’re all doing well! I’m new to the Hailo ecosystem and excited to get started, but I’ve hit a bit of a roadblock and could really use some help.

I recently bought the Raspberry Pi AI HAT+ with the Hailo-8 module and have been trying to get started with running my own models. I installed the Dataflow Compiler and Model Zoo, but I’m struggling to understand how to convert my .onnx model to .hef using the CLI tools.

When I check the help commands (hailomz parse --help or hailomz compile --help), I see an --input model_name parameter, which only allows predefined models from the Model Zoo. But what’s the point of these predefined names if you’re converting your own model? My model is a computer vision model I found online (from GitHub), and it’s not listed there.

Sorry if this is a basic question—I’m new to these topics. Any guidance or pointers would be greatly appreciated!

Thanks!

Welcome to the Hailo Community!

The hailomz command is only for the models in the Model Zoo. You can use the hailo command to do the same for your custom model.

I would recommend using the Hailo AI Software Suite Docker. That makes the whole installation process easier.

To get started I recommend to work trough the tutorials. Inside the Docker call hailo tutorial to start a Jupyter notebook server with notebooks for each individual step.

The CLI will allow you to do some quick test conversions. However using Python scripts is the recommended way to convert models because it offers more flexibility.