Yolo-v5: Optimized for Qualcomm Devices

YoloV5 is a machine learning model that predicts bounding boxes and classes of objects in an image.

This is based on the implementation of Yolo-v5 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-v5 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.object_detection

Model Stats:

• Model checkpoint: YoloV5-M
• Input resolution: 640x640
• Number of parameters: 21.2M
• Model size (float): 81.1 MB
• Model size (w8a16): 21.8 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Yolo-v5	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	5.676 ms	5 - 212 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Elite Mobile	7.465 ms	1 - 203 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® X2 Elite	7.319 ms	47 - 47 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® X Elite	13.713 ms	46 - 46 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® X Elite	13.713 ms	46 - 46 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Gen 3 Mobile	9.62 ms	1 - 248 MB	NPU
Yolo-v5	ONNX	float	Qualcomm® QCS8550 (Proxy)	13.028 ms	0 - 53 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	7.465 ms	1 - 203 MB	NPU
Yolo-v5	ONNX	float	Qualcomm® QCS9075	22.336 ms	5 - 12 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	4.271 ms	3 - 265 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Elite Mobile	5.558 ms	0 - 259 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® X2 Elite	5.072 ms	24 - 24 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® X Elite	10.656 ms	23 - 23 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® X Elite	10.656 ms	23 - 23 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	6.406 ms	3 - 311 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS6490	1715.367 ms	93 - 98 MB	CPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	9.957 ms	0 - 159 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	5.558 ms	0 - 259 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	833.049 ms	95 - 105 MB	CPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCM6690	874.651 ms	90 - 100 MB	CPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS9075	13.676 ms	2 - 5 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	833.049 ms	95 - 105 MB	CPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	4.843 ms	5 - 207 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Elite Mobile	6.482 ms	0 - 199 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® X2 Elite	6.771 ms	5 - 5 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® X Elite	12.102 ms	5 - 5 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® X Elite	12.102 ms	5 - 5 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	8.562 ms	5 - 240 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	11.424 ms	5 - 195 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® SA8775P	17.483 ms	0 - 194 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® SA8775P	17.483 ms	0 - 194 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® SA8775P	17.483 ms	0 - 194 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® SA7255P	63.89 ms	0 - 193 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	25.886 ms	5 - 262 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® SA8295P	22.196 ms	0 - 219 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	6.482 ms	0 - 199 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS9075	18.24 ms	7 - 13 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	3.29 ms	2 - 244 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	4.259 ms	2 - 241 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® X2 Elite	4.373 ms	2 - 2 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® X Elite	9.406 ms	2 - 2 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® X Elite	9.406 ms	2 - 2 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	5.808 ms	2 - 277 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS6490	32.054 ms	3 - 7 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	8.771 ms	2 - 4 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® SA8775P	8.973 ms	1 - 234 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® SA8775P	8.973 ms	1 - 234 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® SA8775P	8.973 ms	1 - 234 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® SA7255P	20.98 ms	0 - 231 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	4.259 ms	2 - 241 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	12.639 ms	2 - 252 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCM6690	130.6 ms	2 - 272 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® SA8295P	12.555 ms	0 - 235 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS9075	10.618 ms	4 - 8 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS8450 (Proxy)	13.454 ms	2 - 280 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	12.639 ms	2 - 252 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	4.463 ms	0 - 233 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Elite Mobile	6.187 ms	0 - 226 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	8.246 ms	0 - 284 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS8550 (Proxy)	10.992 ms	0 - 3 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® SA8775P	17.014 ms	0 - 221 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® SA8775P	17.014 ms	0 - 221 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® SA8775P	17.014 ms	0 - 221 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® SA7255P	63.123 ms	1 - 222 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS8450 (Proxy)	25.464 ms	0 - 311 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® SA8295P	21.612 ms	0 - 248 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	6.187 ms	0 - 226 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS9075	17.992 ms	0 - 57 MB	NPU

License

• The license for the original implementation of Yolo-v5 can be found here.

References

• 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.