RF-DETR: Optimized for Qualcomm Devices

DETR is a machine learning model that can detect objects (trained on COCO dataset).

This is based on the implementation of RF-DETR found here. This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.

Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up to run these models on a hosted Qualcomm® device.

Getting Started

There are two ways to deploy this model on your device:

Option 1: Download Pre-Exported Models

Below are pre-exported model assets ready for deployment.

Runtime	Precision	Chipset	SDK Versions	Download
ONNX	float	Universal	QAIRT 2.45, ONNX Runtime 1.25.0	Download
TFLITE	float	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit RF-DETR on Qualcomm® AI Hub.

Option 2: Export with Custom Configurations

Use the Qualcomm® AI Hub Models Python library to compile and export the model with your own:

• Custom weights (e.g., fine-tuned checkpoints)
• Custom input shapes
• Target device and runtime configurations

This option is ideal if you need to customize the model beyond the default configuration provided here.

See our repository for RF-DETR on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.object_detection

Model Stats:

• Model checkpoint: RF-DETR-base
• Input resolution: 560x560
• Number of parameters: 29.0M
• Model size: 116MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
RF-DETR	ONNX	float	Snapdragon® 8 Gen 3 Mobile	49.945 ms	17 - 687 MB	NPU
RF-DETR	ONNX	float	Snapdragon® 8 Gen 1 Mobile	102.383 ms	17 - 622 MB	NPU
RF-DETR	ONNX	float	Qualcomm® QCS8550 (Proxy)	66.8 ms	0 - 73 MB	NPU
RF-DETR	ONNX	float	Qualcomm® QCS8450	102.383 ms	17 - 622 MB	NPU
RF-DETR	ONNX	float	Snapdragon® 8 Elite Mobile	36.297 ms	13 - 586 MB	NPU
RF-DETR	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	32.012 ms	13 - 614 MB	NPU
RF-DETR	ONNX	float	Qualcomm® QCS9075	83.333 ms	15 - 60 MB	NPU
RF-DETR	ONNX	float	Qualcomm® QCS8750	36.297 ms	13 - 586 MB	NPU
RF-DETR	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	57.076 ms	5 - 756 MB	NPU
RF-DETR	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	118.067 ms	7 - 658 MB	NPU
RF-DETR	TFLITE	float	Qualcomm® QCS8275	171.776 ms	7 - 611 MB	NPU
RF-DETR	TFLITE	float	Qualcomm® QCS8550 (Proxy)	79.714 ms	0 - 4 MB	NPU
RF-DETR	TFLITE	float	Qualcomm® SA8775P	773.946 ms	9 - 25 MB	CPU
RF-DETR	TFLITE	float	Qualcomm® SA8650P	773.946 ms	9 - 25 MB	CPU
RF-DETR	TFLITE	float	Qualcomm® SA8255P	773.946 ms	9 - 25 MB	CPU
RF-DETR	TFLITE	float	Qualcomm® QCS8450	118.067 ms	7 - 658 MB	NPU
RF-DETR	TFLITE	float	Snapdragon® 8 Elite Mobile	42.918 ms	6 - 607 MB	NPU
RF-DETR	TFLITE	float	Qualcomm® SA8295P	105.458 ms	0 - 589 MB	NPU
RF-DETR	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	36.806 ms	6 - 622 MB	NPU
RF-DETR	TFLITE	float	Qualcomm® SA7255P	171.776 ms	7 - 611 MB	NPU
RF-DETR	TFLITE	float	Qualcomm® QCS9075	93.103 ms	6 - 77 MB	NPU
RF-DETR	TFLITE	float	Qualcomm® QCS8750	42.918 ms	6 - 607 MB	NPU

License

• The license for the original implementation of RF-DETR can be found here.

References

• RF-DETR A SOTA Real-Time Object Detection Model
• Source Model Implementation

Community

• Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
• For questions or feedback please reach out to us.