Skip to content

RF-DETR

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.

Supported Model Types

You can deploy the following RF-DETR model types with Inference:

  • Object Detection

Model Overview

Usage Example

You can use RF-DETR with the following code:

import os
import supervision as sv
from inference import get_model
from PIL import Image
from io import BytesIO
import requests

response = requests.get("https://media.roboflow.com/dog.jpeg")

if response.status_code == 200:
    image_data = BytesIO(response.content)

    image = Image.open(image_data)

model = get_model("rfdetr-base")

predictions = model.infer(image, confidence=0.5)[0]

detections = sv.Detections.from_inference(predictions)

labels = [prediction.class_name for prediction in predictions.predictions]

annotated_image = image.copy()
annotated_image = sv.BoxAnnotator().annotate(annotated_image, detections)
annotated_image = sv.LabelAnnotator().annotate(annotated_image, detections, labels)

sv.plot_image(annotated_image)

annotated_image.save("annotated_image_base.jpg")

model = get_model("rfdetr-large")

predictions = model.infer(image, confidence=0.5)[0]

detections = sv.Detections.from_inference(predictions)

labels = [prediction.class_name for prediction in predictions.predictions]

annotated_image = image.copy()
annotated_image = sv.BoxAnnotator().annotate(annotated_image, detections)
annotated_image = sv.LabelAnnotator().annotate(annotated_image, detections, labels)

sv.plot_image(annotated_image)

annotated_image.save("annotated_image_large.jpg")

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

License

See our Licensing Guide for more information about how your use of RF-DETR is licensed when using Inference to deploy your model.