Skip to main content

Model Registration Video Tutorials

Overview​

In this video series we walk through registering a model to Dojo using the MaxHop locust hopper presence prediction model as our use case. This model is written in R but most any model can be registered to Dojo as long as it can be run in a Debian environment. Check out our model registration documentation for more detail.

Getting started​

Model Metadata Capture​

Here we walk through the Dojo model metadata capture process.

Model Installation​

Here we walk through the steps you should take to install your model and verify a successful installation.

Model Configuration​

Here we walk through how to annotate a model configuration file.

Dojo Utilities​

Here we explain the core Dojo utilities that can be leveraged during the installation process.

Model Execution Directive​

Here we explain how to annotate your model's execution command.

Handling Output Files​

Here we describe how to teach Dojo where the model's output files are written and how to make use of them.

Model Accessories​

Here we show how to annotate arbitrary accessory files that your model may produce such as images, maps, or videos.

Finalizing the Model Registration​

Here we recap our process and ensure the model is finalized and sent to Dojo.

Publishing your Model​

The final step in the process is to publish the model. This notifies Dojo's sister system, Causemos, that the model is available for use.