Adding a Driverless AI model to H2O Wave

[GeorgiaJR]

Hello.

I am currently running models on H2O Driverless AI and I want to be able to export the final model and load this into the H2O Wave app that I have running locally on my machine.

I have tried to follow the WaveML tutorials here but these are to build and train models and also require the Wave application to run on the cloud and through the app store.

Is there any way I can export a trained model from Driverless AI and deploy this onto H2O Wave? I am having difficulty finding anything online to explain how to do this.

Thanks in advance!

[sairaamv]

Hi @GeorgiaJR ,

Its a bit open-ended but will give my shot to answer. There are a couple of ways to connect Driverless AI models with H2O Wave.

1. Using H2O AI Cloud as you mentioned
2. Using pip install driverlessai and connect your driverless-ai instance
3. Download Driverless AI model artefacts and upload/connect with H2O Wave

To my exposure, I agree that there is not a straight-up boilerplate available to have these two modules together in a public git repo. I would recommend exploring in automating/downloading model artefacts from driverlessai connector (Path-2) and explore an app which you could modify from any of these available example here.

If possible, put these modules together based on your quick-need and I will definitely PR your repo there.