From the course: MLOps Essentials: Model Deployment and Monitoring
Join today to access over 24,400 courses taught by industry experts.
ML production data best practices
From the course: MLOps Essentials: Model Deployment and Monitoring
ML production data best practices
- [Instructor] As models in production do inference, they also produce significant amounts of valuable data, usually referred to as ML production data. We will discuss some best practices in capturing and managing this data in this video. ML production data contains multiple aspects of inference, usually linked together by specific identifiers. It contains both raw and transformed model inputs. It contains the model outputs or predictions. It also has statistics about the inference, including latency, and any confidence or error measures. True labels may be available in some cases. Sometimes they are manually collected from users. The production data set is valuable for monitoring the performance of the model, as well as serve as new training and testing data for model improvements. However, this is significant data that needs to be captured and processed, and creates significant load on the serving setup. Here are some…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
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.
