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® X2 Elite	7.246 ms	208 - 208 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® X Elite	13.53 ms	144 - 144 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Gen 3 Mobile	9.614 ms	5 - 238 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Gen 1 Mobile	27.883 ms	5 - 264 MB	NPU
Yolo-v5	ONNX	float	Qualcomm® QCS8550 (Proxy)	12.896 ms	0 - 68 MB	NPU
Yolo-v5	ONNX	float	Qualcomm® QCS8450	27.883 ms	5 - 264 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Elite Mobile	7.496 ms	3 - 203 MB	NPU
Yolo-v5	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	5.632 ms	4 - 213 MB	NPU
Yolo-v5	ONNX	float	Qualcomm® QCS9075	21.91 ms	5 - 50 MB	NPU
Yolo-v5	ONNX	float	Qualcomm® QCS8750	7.496 ms	3 - 203 MB	NPU
Yolo-v5	ONNX	float	Qualcomm® QCS7181	13.53 ms	144 - 144 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® X2 Elite	5.05 ms	179 - 179 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® X Elite	10.272 ms	179 - 179 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	6.403 ms	0 - 290 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	14.149 ms	1 - 292 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS6490	35.366 ms	2 - 48 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	9.768 ms	0 - 33 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS8450	14.149 ms	1 - 292 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS9075	13.138 ms	2 - 47 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	4.305 ms	1 - 266 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 8 Elite Mobile	5.548 ms	1 - 262 MB	NPU
Yolo-v5	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	14.346 ms	3 - 267 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCM6690	126.524 ms	0 - 286 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS7790	14.346 ms	3 - 267 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS8750	5.548 ms	1 - 262 MB	NPU
Yolo-v5	ONNX	w8a16	Qualcomm® QCS7181	10.272 ms	179 - 179 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® X2 Elite	6.774 ms	5 - 5 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® X Elite	12.142 ms	5 - 5 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	8.574 ms	0 - 235 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	25.865 ms	5 - 263 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS8275	63.865 ms	1 - 194 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	11.414 ms	5 - 8 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS8450	25.865 ms	5 - 263 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Elite Mobile	6.469 ms	5 - 204 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® SA8295P	22.217 ms	0 - 220 MB	NPU
Yolo-v5	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	4.874 ms	5 - 210 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® SA7255P	63.865 ms	1 - 194 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS9075	18.668 ms	5 - 11 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS8750	6.469 ms	5 - 204 MB	NPU
Yolo-v5	QNN_DLC	float	Qualcomm® QCS7181	12.142 ms	5 - 5 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® X2 Elite	4.341 ms	2 - 2 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® X Elite	9.487 ms	2 - 2 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	5.793 ms	2 - 277 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Gen 1 Mobile	13.583 ms	2 - 281 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS6490	32.347 ms	2 - 6 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS8275	20.985 ms	0 - 232 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	8.728 ms	2 - 5 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS8450	13.583 ms	2 - 281 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS9075	10.646 ms	1 - 5 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	3.304 ms	2 - 248 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	4.246 ms	2 - 242 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® SA8295P	12.56 ms	0 - 235 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	12.571 ms	2 - 252 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCM6690	131.141 ms	2 - 272 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® SA7255P	20.985 ms	0 - 232 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS7790	12.571 ms	2 - 252 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS8750	4.246 ms	2 - 242 MB	NPU
Yolo-v5	QNN_DLC	w8a16	Qualcomm® QCS7181	9.487 ms	2 - 2 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	8.251 ms	0 - 286 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	25.643 ms	0 - 312 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS8275	63.124 ms	0 - 222 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS8550 (Proxy)	10.928 ms	0 - 3 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® SA8775P	108.718 ms	3 - 26 MB	GPU
Yolo-v5	TFLITE	float	Qualcomm® SA8650P	108.718 ms	3 - 26 MB	GPU
Yolo-v5	TFLITE	float	Qualcomm® SA8255P	108.718 ms	3 - 26 MB	GPU
Yolo-v5	TFLITE	float	Qualcomm® QCS8450	25.643 ms	0 - 312 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Elite Mobile	6.177 ms	0 - 227 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® SA8295P	21.644 ms	0 - 250 MB	NPU
Yolo-v5	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	4.478 ms	0 - 234 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® SA7255P	63.124 ms	0 - 222 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS9075	19.056 ms	0 - 57 MB	NPU
Yolo-v5	TFLITE	float	Qualcomm® QCS8750	6.177 ms	0 - 227 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.