Structure for custom entrypoint inference script and its overridden methods for DJLModel

[gdagur]

What did you find confusing? Please describe.
To pass a custom entrypoint inference code different types of framework models support passing a custom script while instantiating them. From what I have observed, for different types of models different methods needs to be overridden in the custom script. For example for HuggingFaceModel script I had to override methods like transform_fn, predict_fn etc, but for TensorflowModel I had to override input_handler, output_handler etc like mentioned here.

Similarly what is the structure looks like of custom entrypoint inference script fo DJLModel

Describe how documentation can be improved
An example in README or in https://github.com/aws/amazon-sagemaker-examples for deploying DJLModel with custom entrypoint script and examples of sample entrypoint script, its structure and overriden methods for DJLModel