README.md

RF-DETR: SOTA Real-Time Object Detection Model

How to Run

1. set up a virtual environment.

   git clone https://github.com/roboflow/rf-detr.git
   cd rf-detr
   conda create -n rfdetr -y python=3.11
   conda activate rfdetr
   pip install torch torchvision --index-url https://download.pytorch.org/whl/cu129
   pip install rfdetr
   pip install opencv-python
   pip install onnx
   pip install onnxscript
   pip install onnxsim
   

2. check pytorch model inference

   cd ..
   python infer.py
   

How to Run (TensorRT)

1. generate onnx file

   python onnx_export_new_ln.py
   

2. generate tensorrt model

   python onnx2trt.py
   

• fp32 (rf_detr_nano)
  [TRT_E] 1000 iterations time: 7.9187 [sec]
  [TRT_E] Average FPS: 126.28 [fps]
  [TRT_E] Average inference time: 7.92 [msec]
  GPU mem : 250M

• fp16 (rf_detr_nano)
  [TRT_E] 1000 iterations time: 3.5675 [sec]
  [TRT_E] Average FPS: 280.31 [fps]
  [TRT_E] Average inference time: 3.57 [msec]
  GPU mem : 190M

• RF-DETR: SOTA Real-Time Object Detection Model

• RF-DETR official GitHub