From the course: Data-Centric Visual AI
Join today to access over 25,500 courses taught by industry experts.
Data sources
- [Instructor] Building better datasets can feel like a dauntless task. The more, the more you put into it, it just feels like you're never getting over the hill. However, we're going to talk today about how you can find the best data available for you and hopefully make that feel like a weightless task for you whenever you're building new datasets. After all, the villain of ML is bad data, whether it's model bias issues, lethal physical danger, or just an overall reduction in your model performance, having bad data is going to ultimately be one of the main causes your model doesn't make it to production. Data can be found anywhere, and whether it's at popular AI sites, like Huggingface or Kaggle, videos and images collected elsewhere online, or even synthetically creating your own, finding data for your next visual AI project can be easy, but finding the right data at times can be very difficult. We'll be covering three main methods of collecting data, open-source datasets, scraping…
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.