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:

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

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 For Galaxy Mobile	4.003 ms	3 - 181 MB	NPU
Yolo-v6	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	4.387 ms	0 - 176 MB	NPU
Yolo-v6	ONNX	float	Snapdragon® X2 Elite	3.911 ms	5 - 5 MB	NPU
Yolo-v6	ONNX	float	Snapdragon® X Elite	8.633 ms	8 - 8 MB	NPU
Yolo-v6	ONNX	float	Snapdragon® 8 Gen 3 Mobile	5.209 ms	5 - 210 MB	NPU
Yolo-v6	ONNX	float	Snapdragon® 8 Gen 1 Mobile	11.484 ms	5 - 213 MB	NPU
Yolo-v6	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	9.064 ms	5 - 9 MB	NPU
Yolo-v6	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	8.417 ms	0 - 11 MB	NPU
Yolo-v6	ONNX	float	Qualcomm® QCS8450	11.484 ms	5 - 213 MB	NPU
Yolo-v6	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	10.012 ms	5 - 8 MB	NPU
Yolo-v6	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	8.633 ms	8 - 8 MB	NPU
Yolo-v6	ONNX	float	Qualcomm® Dragonwing™ Q-8750	4.387 ms	0 - 176 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	3.296 ms	5 - 174 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	3.929 ms	0 - 167 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® X2 Elite	3.733 ms	5 - 5 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® X Elite	6.979 ms	5 - 5 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	4.247 ms	5 - 194 MB	NPU
Yolo-v6	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	9.431 ms	5 - 196 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	7.973 ms	5 - 12 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	6.436 ms	5 - 65 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® QCS8450	9.431 ms	5 - 196 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	8.14 ms	5 - 12 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	6.979 ms	5 - 5 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	3.929 ms	0 - 167 MB	NPU
Yolo-v6	QNN_DLC	float	Qualcomm® SA8295P	9.196 ms	0 - 163 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	1.348 ms	0 - 68 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	1.731 ms	2 - 67 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® X2 Elite	1.814 ms	2 - 2 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® X Elite	3.982 ms	2 - 2 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	2.3 ms	0 - 86 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Snapdragon® 8 Gen 1 Mobile	4.948 ms	2 - 93 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	3.43 ms	2 - 7 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	3.573 ms	2 - 73 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® SA8650P	4.133 ms	1 - 67 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® SA8255P	4.133 ms	1 - 67 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® QCS8450	4.948 ms	2 - 93 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	3.891 ms	1 - 6 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	3.982 ms	2 - 2 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	1.731 ms	2 - 67 MB	NPU
Yolo-v6	QNN_DLC	w8a16	Qualcomm® SA8295P	5.483 ms	0 - 65 MB	NPU
Yolo-v6	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	3.268 ms	0 - 178 MB	NPU
Yolo-v6	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	3.985 ms	0 - 170 MB	NPU
Yolo-v6	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	4.374 ms	0 - 202 MB	NPU
Yolo-v6	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	9.427 ms	0 - 201 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	8.004 ms	0 - 21 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	6.412 ms	0 - 4 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® SA8775P	8.297 ms	0 - 173 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® SA8650P	8.297 ms	0 - 173 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® SA8255P	8.297 ms	0 - 173 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® QCS8450	9.427 ms	0 - 201 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	8.141 ms	0 - 21 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	3.985 ms	0 - 170 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® SA7255P	18.356 ms	0 - 171 MB	NPU
Yolo-v6	TFLITE	float	Qualcomm® SA8295P	9.459 ms	0 - 171 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.