Model Testing

To facilitate model development for the Greenhub.ai platform, we provide a command-line tool that is installed alongside the Greenhub SDK.

With this Greenhub command-line tool, you can check whether a given directory meets the requirements for being zipped and uploaded to the Greenhub platform, reducing the chances of errors occurring during the model upload later on.

Hint: More about hosting a model on Greenhub.ai.

Typical procedure

Once you have packaged your model in a run() method within your run.py, you can use the greenhub test in you command line in the directory of the run.py file, to verify if the run file and the run method meet the requirements for a model upload. If an error occurs, a detailed error message will be returned, clarifying what is still wrong.

cd /path/to/model/implementation/directory

greenhub test