Summary
In this chapter, we discussed Ray Serve. Ray Serve is a tool with a massive amount of features. Our goal was not to introduce everything about this tool in this chapter. Rather, we wanted to give you an introduction to the tool and outline the basic requirements to understand how this tool can be used in serving production-ready models while following different patterns of the serving model.
We provided an introduction to the key concepts of the tools, along with examples. We then used Ray Serve to serve two dummy models from end-to-end using the ensemble pattern and the pipeline pattern. Using these examples, we saw how Ray Serve can help you set up model serving while following the standard patterns of serving ML models.
In the next chapter, we will introduce you to BentoML, another tool for serving ML models.
