MLServer Issue: Multiple Model Loads

Reproduction Steps:

0. Install the requirements:

   pip install -r requirements.txt

1. Start the server:

   mlserver start .

   The logs should look like this:

   --------- Model Readiness: True ---------
   INFO:     Started server process [12556]
   INFO:     Waiting for application startup.
   INFO:     Application startup complete.
   INFO:     Uvicorn running on http://0.0.0.0:8080 (Press CTRL+C to quit)

2. Test the server in a seperate terminal:

   python test.py

   The test script logs should indicate inference success:

   {'model_name': 'mlserver-example', 'model_version': None, 'id': '9ca77367-83a1-43e9-884d-3427bad82bad', 'parameters': None, 'outputs': [{'name': 'image_tags', 'shape': [4], 'datatype': 'int', 'parameters': None, 'data': [14.4, 9.6, 28.799999999999997, -2.4]}]}

   But the server logs should indicate that the load() method is being called multiple times unnecessarily:

   --------- Model Readiness: True ---------
   INFO:     127.0.0.1:61218 - "POST /v2/models/mlserver-example/infer HTTP/1.1" 200 OK
   --------- Model Readiness: True ---------
   --------- Model Readiness: True ---------
   --------- Model Readiness: True ---------