Yolo-R: Optimized for Qualcomm Devices

YoloR is a machine learning model that predicts bounding boxes and classes of objects in an image.

This is based on the implementation of Yolo-R 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

Due to licensing restrictions, we cannot distribute pre-exported model assets for this model. 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

See our repository for Yolo-R on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.object_detection

Model Stats:

• Model checkpoint: yolor_p6
• Input resolution: 640x640
• Number of parameters: 4.68M
• Model size (float): 17.9 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Yolo-R	ONNX	float	Snapdragon® X Elite	58.077 ms	74 - 74 MB	NPU
Yolo-R	ONNX	float	Snapdragon® 8 Gen 3 Mobile	41.792 ms	6 - 352 MB	NPU
Yolo-R	ONNX	float	Qualcomm® QCS8550 (Proxy)	56.438 ms	0 - 79 MB	NPU
Yolo-R	ONNX	float	Qualcomm® QCS9075	52.883 ms	5 - 12 MB	NPU
Yolo-R	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	37.174 ms	3 - 234 MB	NPU
Yolo-R	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	26.905 ms	6 - 318 MB	NPU
Yolo-R	ONNX	float	Snapdragon® X2 Elite	26.205 ms	75 - 75 MB	NPU
Yolo-R	ONNX	w8a16	Snapdragon® X Elite	30.372 ms	40 - 40 MB	NPU
Yolo-R	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	20.883 ms	2 - 498 MB	NPU
Yolo-R	ONNX	w8a16	Qualcomm® QCS6490	2267.909 ms	127 - 136 MB	CPU
Yolo-R	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	29.364 ms	0 - 49 MB	NPU
Yolo-R	ONNX	w8a16	Qualcomm® QCS9075	30.707 ms	1 - 6 MB	NPU
Yolo-R	ONNX	w8a16	Qualcomm® QCM6690	1177.508 ms	80 - 93 MB	CPU
Yolo-R	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	17.123 ms	1 - 354 MB	NPU
Yolo-R	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	1121.096 ms	125 - 138 MB	CPU
Yolo-R	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	17.966 ms	3 - 428 MB	NPU
Yolo-R	ONNX	w8a16	Snapdragon® X2 Elite	18.3 ms	41 - 41 MB	NPU
Yolo-R	QNN_DLC	float	Snapdragon® X Elite	29.418 ms	5 - 5 MB	NPU
Yolo-R	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	22.571 ms	2 - 322 MB	NPU
Yolo-R	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	100.259 ms	1 - 253 MB	NPU
Yolo-R	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	29.695 ms	5 - 40 MB	NPU
Yolo-R	QNN_DLC	float	Qualcomm® SA8775P	36.081 ms	1 - 259 MB	NPU
Yolo-R	QNN_DLC	float	Qualcomm® QCS9075	37.318 ms	5 - 11 MB	NPU
Yolo-R	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	50.214 ms	4 - 377 MB	NPU
Yolo-R	QNN_DLC	float	Qualcomm® SA7255P	100.259 ms	1 - 253 MB	NPU
Yolo-R	QNN_DLC	float	Qualcomm® SA8295P	44.061 ms	0 - 307 MB	NPU
Yolo-R	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	18.489 ms	0 - 232 MB	NPU
Yolo-R	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	14.036 ms	3 - 294 MB	NPU
Yolo-R	QNN_DLC	float	Snapdragon® X2 Elite	15.22 ms	5 - 5 MB	NPU
Yolo-R	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	39.226 ms	0 - 442 MB	NPU
Yolo-R	TFLITE	float	Qualcomm® QCS8275 (Proxy)	137.603 ms	1 - 328 MB	NPU
Yolo-R	TFLITE	float	Qualcomm® QCS8550 (Proxy)	59.157 ms	1 - 3 MB	NPU
Yolo-R	TFLITE	float	Qualcomm® SA8775P	66.396 ms	1 - 310 MB	NPU
Yolo-R	TFLITE	float	Qualcomm® QCS9075	50.958 ms	1 - 86 MB	NPU
Yolo-R	TFLITE	float	Qualcomm® QCS8450 (Proxy)	88.808 ms	0 - 468 MB	NPU
Yolo-R	TFLITE	float	Qualcomm® SA7255P	137.603 ms	1 - 328 MB	NPU
Yolo-R	TFLITE	float	Qualcomm® SA8295P	76.893 ms	1 - 355 MB	NPU
Yolo-R	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	31.724 ms	1 - 288 MB	NPU
Yolo-R	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	26.427 ms	1 - 342 MB	NPU

License

• The license for the original implementation of Yolo-R can be found here.

References

• You Only Learn One Representation: Unified Network for Multiple Tasks
• 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.