// ◉ Multimodal & Vision
DETR (Detection Transformer)
Casts detection as set prediction with a Transformer – no anchors, no NMS.
DETR（Transformer 目标检测） 2020 Intermediate core
IN PLAIN WORDS
A CNN extracts features, a Transformer turns a fixed set of object queries into boxes (or “no object”), and bipartite matching pairs predictions with ground truth during training. Hand-crafted anchors and non-maximum suppression disappear, opening the line of Transformer-based detectors.
Used in: The basis for Deformable DETR, RT-DETR and open-vocabulary detectors
TIMELINE · prerequisites → this → what builds on it
01Original paper
End-to-End Object Detection with Transformers
Nicolas Carion, Francisco Massa, Gabriel Synnaeve, et al. · 2020
arXiv ↗ ✓ Verified · arXiv API
02Where it sits
Learn the left column first; the right column builds on this one.
Then you can learn
- Nothing in the atlas builds on it yet
Related:Vision Transformer (ViT)Segment Anything (SAM)
Left: prerequisites · right: builds on this · dashed: related. Full map →
03Learning resources
Official docs, explainers, courses and reference code. All links checked.
04Where the course teaches it
For AI / data
- Verification: verified, checked 2026-10-22.
- JSON: /api/learn/algorithms.json (id:
detr)