May 2020
Intermediate to advanced
404 pages
10h 52m
English
We have a model that is trained by a separate script on the backend and is stored as a model and then deployed as an API-based service. Here, we're looking at a solution that produces results on-demand but the training occurs offline (not in the execution span of the portion of code responsible for responding to the client queries). Web APIs respond to a single query at a time and yield singular results.
This is by far the most commonly used method for deploying DL in production since it allows accurate training performed offline by data scientists and a short deployment script to create an API. In this book, we have mostly carried out deployments of this kind.
Read now
Unlock full access