Video Highlights: Scalable Object Detection with DETR

Print Friendly, PDF & Email

Object detection is a central problem in computer vision and underpins many applications from medical image analysis to autonomous driving. This video presentation will start with a tutorial on object detection covering basic concepts and techniques. Then we will dive into an interactive session where you will implement a recent method introduced by Facebook called DEtection TRansformers (DETR) and learn how to use Determined’s integrated deep learning platform to easily train models for object detection at scale.

The presenter is Liam Li, a ML Engineer at Determined AI. Previously, he received his Ph.D. in machine learning at CMU where he focused on AutoML.

Key Takeaways:

  • Basics of object detection including main concepts and techniques
  • Guided implementation of Facebook’s DETR model
  • Hands-on tour of the core capabilities of Determined, including experiment management, distributed training, and hyperparameter search

The slides for the presentation are available HERE. And the code is available HERE.

Sign up for the free insideBIGDATA newsletter.

Join us on Twitter: @InsideBigData1 –

Speak Your Mind