Great overview on making the model production ready and thanks for pointing out FastAPI 🙏

Any reason/advantage to use FastAPI instead of TensorFlow serving or TorchServe in case of Deep Learning models? Or do you advice to use both together?