aim-uofa/AdelaiDet
A PyTorch-based open source toolbox for instance-level detection and recognition tasks including object detection, instance segmentation, and text detection.

Velocity · 7d
+1.5
★ / day
Trend
→steady
star history
AdelaiDet provides implementations of multiple computer vision algorithms built on top of Detectron2. It includes models for object detection (FCOS), instance segmentation (BlendMask, CondInst, SOLO, SOLOv2, BoxInst), and text detection/recognition (ABCNet, ABCNetv2). The toolbox offers pre-trained model weights hosted on HuggingFace and benchmarks on standard datasets like COCO.