From the course: Data-Centric Visual AI
Join today to access over 25,500 courses taught by industry experts.
Developing competing models
From the course: Data-Centric Visual AI
Developing competing models
- [Instructor] Welcome to the last notebook in the course where we're going to be looking at developing competing models. One of the most critical portions of developing any model lifecycle or model production line is that at some point you're going to have a new model that is going to eventually usurp the throne from your previous model. But how do you know when that new model is good enough to push into production? We're going to look at exactly how you can figure out when to push a new model production by using FiftyOne, as well as some additional tools to help us gain those insights as well. So let's dive in. First with installation, we're going to be using FiftyOne umap learn scikit learn as well as ultralytics to help us find these differences in our dataset. And when it comes to loading our dataset, we're going to be using that same VisDrone dataset again. However, in this case, we've added an extra model as well as some evaluations as well to the dataset. Note the difference…
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.