Yolo-v6: Optimized for Mobile Deployment

Real-time object detection optimized for mobile and edge

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

This model is an implementation of Yolo-v6 found here.

More details on model performance across various devices, can be found here.

Model Details

Model Type: Object detection
Model Stats:
- Model checkpoint: YoloV6-N
- Input resolution: 640x640
- Number of parameters: 4.68M
- Model size: 17.9 MB

Model	Device	Chipset	Target Runtime	Inference Time (ms)	Peak Memory Range (MB)	Precision	Primary Compute Unit	Target Model
Yolo-v6	Samsung Galaxy S23	Snapdragon® 8 Gen 2	TFLITE	10.657 ms	0 - 9 MB	FP16	NPU	--
Yolo-v6	Samsung Galaxy S23	Snapdragon® 8 Gen 2	QNN	5.675 ms	5 - 17 MB	FP16	NPU	--
Yolo-v6	Samsung Galaxy S23	Snapdragon® 8 Gen 2	ONNX	5.507 ms	5 - 37 MB	FP16	NPU	--
Yolo-v6	Samsung Galaxy S24	Snapdragon® 8 Gen 3	TFLITE	7.294 ms	0 - 49 MB	FP16	NPU	--
Yolo-v6	Samsung Galaxy S24	Snapdragon® 8 Gen 3	QNN	3.962 ms	5 - 48 MB	FP16	NPU	--
Yolo-v6	Samsung Galaxy S24	Snapdragon® 8 Gen 3	ONNX	4.222 ms	5 - 70 MB	FP16	NPU	--
Yolo-v6	Snapdragon 8 Elite QRD	Snapdragon® 8 Elite	TFLITE	7.206 ms	0 - 43 MB	FP16	NPU	--
Yolo-v6	Snapdragon 8 Elite QRD	Snapdragon® 8 Elite	QNN	4.214 ms	5 - 39 MB	FP16	NPU	--
Yolo-v6	Snapdragon 8 Elite QRD	Snapdragon® 8 Elite	ONNX	4.302 ms	5 - 55 MB	FP16	NPU	--
Yolo-v6	SA7255P ADP	SA7255P	TFLITE	85.024 ms	0 - 38 MB	FP16	NPU	--
Yolo-v6	SA7255P ADP	SA7255P	QNN	78.113 ms	2 - 11 MB	FP16	NPU	--
Yolo-v6	SA8255 (Proxy)	SA8255P Proxy	TFLITE	10.574 ms	0 - 11 MB	FP16	NPU	--
Yolo-v6	SA8255 (Proxy)	SA8255P Proxy	QNN	4.6 ms	5 - 7 MB	FP16	NPU	--
Yolo-v6	SA8295P ADP	SA8295P	TFLITE	13.93 ms	0 - 29 MB	FP16	NPU	--
Yolo-v6	SA8295P ADP	SA8295P	QNN	7.081 ms	0 - 18 MB	FP16	NPU	--
Yolo-v6	SA8650 (Proxy)	SA8650P Proxy	TFLITE	10.602 ms	0 - 8 MB	FP16	NPU	--
Yolo-v6	SA8650 (Proxy)	SA8650P Proxy	QNN	4.746 ms	5 - 7 MB	FP16	NPU	--
Yolo-v6	SA8775P ADP	SA8775P	TFLITE	14.1 ms	0 - 37 MB	FP16	NPU	--
Yolo-v6	SA8775P ADP	SA8775P	QNN	7.424 ms	0 - 10 MB	FP16	NPU	--
Yolo-v6	QCS8275 (Proxy)	QCS8275 Proxy	TFLITE	85.024 ms	0 - 38 MB	FP16	NPU	--
Yolo-v6	QCS8275 (Proxy)	QCS8275 Proxy	QNN	78.113 ms	2 - 11 MB	FP16	NPU	--
Yolo-v6	QCS8550 (Proxy)	QCS8550 Proxy	TFLITE	10.491 ms	0 - 8 MB	FP16	NPU	--
Yolo-v6	QCS8550 (Proxy)	QCS8550 Proxy	QNN	4.538 ms	5 - 7 MB	FP16	NPU	--
Yolo-v6	QCS9075 (Proxy)	QCS9075 Proxy	TFLITE	14.1 ms	0 - 37 MB	FP16	NPU	--
Yolo-v6	QCS9075 (Proxy)	QCS9075 Proxy	QNN	7.424 ms	0 - 10 MB	FP16	NPU	--
Yolo-v6	QCS8450 (Proxy)	QCS8450 Proxy	TFLITE	13.048 ms	0 - 39 MB	FP16	NPU	--
Yolo-v6	QCS8450 (Proxy)	QCS8450 Proxy	QNN	7.726 ms	5 - 32 MB	FP16	NPU	--
Yolo-v6	Snapdragon X Elite CRD	Snapdragon® X Elite	QNN	4.842 ms	5 - 5 MB	FP16	NPU	--
Yolo-v6	Snapdragon X Elite CRD	Snapdragon® X Elite	ONNX	6.095 ms	7 - 7 MB	FP16	NPU	--

License

The license for the original implementation of Yolo-v6 can be found here.
The license for the compiled assets for on-device deployment can be found here

References

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.

Usage and Limitations

Model may not be used for or in connection with any of the following applications:

Accessing essential private and public services and benefits;
Administration of justice and democratic processes;
Assessing or recognizing the emotional state of a person;
Biometric and biometrics-based systems, including categorization of persons based on sensitive characteristics;
Education and vocational training;
Employment and workers management;
Exploitation of the vulnerabilities of persons resulting in harmful behavior;
General purpose social scoring;
Law enforcement;
Management and operation of critical infrastructure;
Migration, asylum and border control management;
Predictive policing;
Real-time remote biometric identification in public spaces;
Recommender systems of social media platforms;
Scraping of facial images (from the internet or otherwise); and/or
Subliminal manipulation