Yolo-v6: Optimized for Qualcomm Devices

YoloV6 is a machine learning model that predicts bounding boxes and classes of objects in an image.

This is based on the implementation of Yolo-v6 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-v6 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.object_detection

Model Stats:

• Model checkpoint: YoloV6-N
• Input resolution: 640x640
• Number of parameters: 4.68M
• Model size (float): 17.9 MB
• Model size (w8a8): 4.68 MB
• Model size (w8a16): 5.03 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Yolo-v6	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	3.997 ms	5 - 172 MB	NPU
Yolo-v6	ONNX	float	Snapdragon® X2 Elite	3.678 ms	14 - 14 MB	NPU
Yolo-v6	ONNX	float	Snapdragon® X Elite	8.342 ms	14 - 14 MB	NPU
Yolo-v6	ONNX	float	Snapdragon® 8 Gen 3 Mobile	5.5 ms	5 - 203 MB	NPU
Yolo-v6	ONNX	float	Qualcomm® QCS8550 (Proxy)	8.111 ms	0 - 17 MB	NPU
Yolo-v6	ONNX	float	Qualcomm® QCS9075	9.655 ms	5 - 7 MB	NPU
Yolo-v6	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	4.317 ms	1 - 170 MB	NPU
Yolo-v6	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	1.841 ms	0 - 187 MB	NPU
Yolo-v6	ONNX	w8a16	Snapdragon® X2 Elite	2.009 ms	5 - 5 MB	NPU
Yolo-v6	ONNX	w8a16	Snapdragon® X Elite	4.529 ms	3 - 3 MB	NPU
Yolo-v6	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	2.677 ms	0 - 221 MB	NPU
Yolo-v6	ONNX	w8a16	Qualcomm® QCS6490	286.42 ms	40 - 45 MB	CPU
Yolo-v6	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	4.08 ms	0 - 7 MB	NPU
Yolo-v6	ONNX	w8a16	Qualcomm® QCS9075	4.858 ms	2 - 5 MB	NPU
Yolo-v6	ONNX	w8a16	Qualcomm® QCM6690	147.177 ms	42 - 50 MB	CPU
Yolo-v6	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	2.079 ms	0 - 181 MB	NPU
Yolo-v6	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	128.373 ms	44 - 52 MB	CPU
Yolo-v6	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	3.113 ms	5 - 161 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® X2 Elite	3.205 ms	5 - 5 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® X Elite	6.244 ms	5 - 5 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	4.482 ms	0 - 182 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	16.205 ms	0 - 153 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	6.139 ms	5 - 19 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® SA8775P	7.67 ms	0 - 157 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® QCS9075	7.742 ms	7 - 13 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	8.901 ms	5 - 188 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® SA7255P	16.205 ms	0 - 153 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® SA8295P	9.081 ms	3 - 155 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	3.326 ms	0 - 156 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	0.949 ms	0 - 40 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® X2 Elite	1.403 ms	2 - 2 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® X Elite	2.492 ms	2 - 2 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	1.503 ms	2 - 57 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® QCS6490	6.586 ms	2 - 6 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® QCS8275 (Proxy)	5.353 ms	1 - 38 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	2.205 ms	2 - 3 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® SA8775P	2.819 ms	1 - 41 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® QCS9075	2.58 ms	1 - 5 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® QCM6690	17.547 ms	2 - 152 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® QCS8450 (Proxy)	2.817 ms	2 - 58 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® SA7255P	5.353 ms	1 - 38 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® SA8295P	3.478 ms	2 - 37 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	1.15 ms	2 - 48 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	2.735 ms	0 - 150 MB	NPU
Yolo-v6	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	3.389 ms	0 - 166 MB	NPU
Yolo-v6	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	11.18 ms	4 - 77 MB	GPU
Yolo-v6	TFLITE	float	Qualcomm® QCS8275 (Proxy)	56.698 ms	6 - 35 MB	GPU
Yolo-v6	TFLITE	float	Qualcomm® QCS8550 (Proxy)	14.579 ms	5 - 24 MB	GPU
Yolo-v6	TFLITE	float	Qualcomm® SA8775P	23.976 ms	6 - 61 MB	GPU
Yolo-v6	TFLITE	float	Qualcomm® QCS9075	7.896 ms	0 - 18 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® QCS8450 (Proxy)	18.4 ms	6 - 84 MB	GPU
Yolo-v6	TFLITE	float	Qualcomm® SA7255P	56.698 ms	6 - 35 MB	GPU
Yolo-v6	TFLITE	float	Qualcomm® SA8295P	19.09 ms	5 - 61 MB	GPU
Yolo-v6	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	3.768 ms	0 - 163 MB	NPU

License

• The license for the original implementation of Yolo-v6 can be found here.

References

• YOLOv6: A Single-Stage Object Detection Framework for Industrial Applications
• 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.