Debugging Guide

You need to train your model before invoking it. Click Train to load the model into memory before making an invocation.

Make sure your Python dependencies are added to the Dependencies section. If you need to customize your container, you can use setup.sh, but please try to install dependencies in the dependencies section first.

Make sure your request includes the fields you defined in handler.py. The handler will enforce whichever fields you specify - if your handler expects to receive a string field, it will return a 400 if you pass in a float.

This can happen when the slai-cli SDK is outdated. Run pip install --upgrade slai-cli to upgrade to the latest version.