RF-DETR is a real-time object detection transformer-based architecture designed to transfer well to both a wide variety of domains and to datasets big and small.
RF-DETR is the first real-time model to exceed 60 AP on the Microsoft COCO benchmark alongside competitive performance at base sizes. It also achieves state-of-the-art performance on RF100-VL, an object detection benchmark that measures model domain adaptability to real world problems. RF-DETR is comparable speed to current real-time objection models.
The model comes in two variants:
RF-DETR Base, which has 29M parameters, and;
RF-DETR Large, which has 129M parameters.
The RF-DETR source code and COCO checkpoint weights is available under an Apache 2.0 license.
If you have fine-tuned an RF-DETR model on Roboflow, you can deploy it by replacing the rfdetr-base model ID in the get_model() function call above with the ID of your trained model on Roboflow. This model ID will look like model-name/1, where model-name is the name of the model and 1 is your model version. Learn how to find your model ID