From the course: MLOps with Databricks
Join today to access over 24,900 courses taught by industry experts.
When to use Feature Serving vs. Model Serving - Databricks Tutorial
From the course: MLOps with Databricks
When to use Feature Serving vs. Model Serving
- [Instructor] Many data science teams struggle to bring machine learning models to production when the serving component is involved. Databricks provides serving capabilities that help those teams simplify their deployment process. In this video, I discussed two architectures, batch serving and real-time model serving, and how both scenarios can be implemented on Databricks. In case of batch serving, predictions are computed in advance using a batch process, stored in SQL or in-memory database, and retrieve at the request. The batch serving architecture is very popular in the case of personal recommendation with low latency requirements. For example, an e-commerce store may recommend products to customers on various pages of the website. In the case of real-time model serving, predictions are computed at the moment a user requests it. To compute the predictions, a model artifact is required. Model artifact is created in a batch fashion as the result of a model training pipeline…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.