Grounding DINO 1.5

model paper
IDEA Lab
Advanced open-set object detection suite with Pro and Edge variants. Pro scales up architecture with enhanced vision backbone and 20M+ training images, achieving 54.3 AP on COCO and 55.7 AP on LVIS-minival zero-shot. Edge is optimized for deployment at 75.2 FPS with TensorRT.
Paper (arXiv)GitHub (API)

Outputs 2

Grounding DINO 1.5

model

Grounding DINO 1.5 Pro for high-accuracy open-set detection and Grounding DINO 1.5 Edge for efficient edge deployment.

GitHub (API)

Variants

Name Parameters Notes
Grounding DINO 1.5 Pro — High-performance, strong generalization, 54.3 AP on COCO
Grounding DINO 1.5 Edge — Optimized for edge deployment, 75.2 FPS with TensorRT

Grounding DINO 1.5: Advance the Edge of Open-Set Object Detection

paper

Technical report on scaling Grounding DINO with larger architecture, enhanced backbone, and expanded training data.

Paper (arXiv)

arXiv: 2405.10300

visionopen-vocabulary