Sagemaker MultiModel Endpoint Custom Entrypoint Script

Question

Im working on a multimodel endpoint for a few tensorflow models leveraging this walkthrough:

https://github.com/aws-samples/sagemaker-multi-model-endpoint-tensorflow-computer-vision/blob/main/multi-model-endpoint-tensorflow-cv.ipynb

My code looks something like this:

env = {
    'SAGEMAKER_REQUIREMENTS': 'requirements.txt', # path relative to `source_dir` below.
    'CATEGORY': category
}

model_1 = TensorFlowModel(model_data=model_loc,
                          role=sagemaker_role,
                          framework_version="2.2.0",
                          source_dir="<path to source dir>",
                          entry_point="<source dir>/inference.py",
                          env=env,
                          sagemaker_session=sagemaker_session)

mme = MultiDataModel(name='test-endpoint',
                     model_data_prefix=model_prefix,
                     model=model_1,
                     sagemaker_session=sagemaker_session)

predictor = mme.deploy(instance_type="ml.r5.xlarge", 
                       initial_instance_count=1
                      )

But when I try to invoke this endpoint, the error returned makes it seem like my custom entry point script isn't referenced at all? This works as intended for a single model endpoint. Any examples or documentation for a multi-model endpoint + custom entry point? I can't find much.