Measure scalability (more than 2 models)

**Intro**

`FastAPI` provides **native async support** (based on Starlette + Uvicorn), which means we can scale more efficiently the number of models (2 in this case, but we could add some more models to get a further improvement).

**Proposal**

Use the number of models as an input argument to Scoring handler API.

So, we can get more information about the scalability if we run different numbers of models from `Locust` (load testing tool).