RF-DETR
Roboflow's real-time object detection — first 60+ mAP real-time detector on COCO
Roboflow's real-time object detection model; the first 60+ mAP real-time detector on COCO.
Image generation and computer vision — detection, editing, 3D.
Key features
- Real-time transformer for object detection, instance segmentation, and keypoint
- Built on DINOv2 vision transformer backbone with state-of-the-art accuracy-latency
- Single consistent API for detection, segmentation, and keypoint tasks
- Models created with neural architecture search; sizes from N to 2XL with in-house COCO
- Open-source package released under Apache 2.0; install via pip install rfdetr
Good for
- Real-time object detection and instance segmentation on video
- Robotics and edge applications needing fast transformer-based
Stack & tags
Pythonobject-detectioninstance-segmentationrealtimetransformerdinov2roboflowapache-2.0
Quick start
pip install rfdetrFrom the project README on GitHub.
Similar open-source projects
FLUX.2 klein
★ 25,996Black Forest Labs' sub-second image generation — generate and edit in <0.5s
DiffSynth-Studio
★ 13,200The framework behind Qwen-Image-i2L — the first "Image→LoRA" visual generation system