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® X Elite	13.554 ms	46 - 46 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Gen 3 Mobile	9.662 ms	1 - 281 MB	NPU
Yolo-v5	ONNX	float	Qualcomm® QCS8550 (Proxy)	12.862 ms	0 - 54 MB	NPU
Yolo-v5	ONNX	float	Qualcomm® QCS9075	22.415 ms	5 - 12 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	7.321 ms	2 - 233 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	5.291 ms	5 - 216 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® X2 Elite	11.001 ms	46 - 46 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® X Elite	10.287 ms	23 - 23 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	6.166 ms	3 - 311 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS6490	1704.695 ms	93 - 96 MB	CPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	9.591 ms	0 - 29 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS9075	12.731 ms	2 - 5 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCM6690	871.082 ms	84 - 94 MB	CPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	5.372 ms	0 - 259 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	834.105 ms	100 - 111 MB	CPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	4.08 ms	3 - 270 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® X2 Elite	4.9 ms	24 - 24 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® X Elite	11.926 ms	5 - 5 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	8.467 ms	0 - 235 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	63.683 ms	1 - 189 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	11.641 ms	5 - 6 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS9075	17.796 ms	5 - 11 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	25.748 ms	5 - 261 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	6.374 ms	0 - 197 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	4.777 ms	5 - 208 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® X2 Elite	6.704 ms	5 - 5 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® X Elite	9.542 ms	2 - 2 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	5.854 ms	2 - 286 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS6490	31.999 ms	4 - 8 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS8275 (Proxy)	21.058 ms	1 - 232 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	8.821 ms	2 - 4 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS9075	10.973 ms	2 - 6 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCM6690	98.601 ms	2 - 268 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS8450 (Proxy)	13.484 ms	2 - 285 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	4.29 ms	2 - 238 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	13.047 ms	2 - 251 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	3.215 ms	2 - 246 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® X2 Elite	4.261 ms	2 - 2 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	8.196 ms	0 - 290 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS8275 (Proxy)	63.131 ms	0 - 222 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS8550 (Proxy)	10.363 ms	0 - 3 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS9075	17.997 ms	0 - 57 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS8450 (Proxy)	30.628 ms	0 - 315 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	6.078 ms	0 - 227 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	4.43 ms	0 - 235 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.