Multi-Task Computer Vision
Performs object detection, instance segmentation, and image classification on images and videos in real-time.
Multiple Export Formats
Exports models to ONNX, CoreML, TFLite, TensorFlow, and OpenCV DNN formats for deployment across platforms.
PyTorch Hub Integration
Supports inference via PyTorch Hub with automatic pretrained model downloads.
Diverse Input and Output Support
Accepts inputs like URLs, PIL images, OpenCV images, NumPy arrays, and PyTorch tensors; outputs detections in torch, pandas, or JSON formats.
Model Variants
Includes lightweight Nano models (YOLOv5n) and instance segmentation models up to version 7.0.