GPUNet: Optimized for Qualcomm Devices

GPUNet is a machine learning model that can classify images from the Imagenet dataset. It can also be used as a backbone in building more complex models for specific use cases.

This is based on the implementation of GPUNet 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

There are two ways to deploy this model on your device:

Option 1: Download Pre-Exported Models

Below are pre-exported model assets ready for deployment.

Runtime	Precision	Chipset	SDK Versions	Download
ONNX	float	Universal	QAIRT 2.42, ONNX Runtime 1.24.3	Download
ONNX	w8a16	Universal	QAIRT 2.42, ONNX Runtime 1.24.3	Download
ONNX	w8a8	Universal	QAIRT 2.42, ONNX Runtime 1.24.3	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a16	Universal	QAIRT 2.45	Download
QNN_DLC	w8a8	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download
TFLITE	w8a8	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit GPUNet on Qualcomm® AI Hub.

Option 2: Export with Custom Configurations

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

This option is ideal if you need to customize the model beyond the default configuration provided here.

See our repository for GPUNet on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

• Model checkpoint: Imagenet
• Input resolution: 224x224
• Number of parameters: 10.49M
• Model size (float): 45.28MB
• Model size (w8a8): 21.3MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
GPUNet	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	0.503 ms	0 - 36 MB	NPU
GPUNet	ONNX	float	Snapdragon® 8 Elite Mobile	0.605 ms	0 - 36 MB	NPU
GPUNet	ONNX	float	Snapdragon® X2 Elite	0.483 ms	24 - 24 MB	NPU
GPUNet	ONNX	float	Snapdragon® X Elite	1.149 ms	24 - 24 MB	NPU
GPUNet	ONNX	float	Snapdragon® X Elite	1.149 ms	24 - 24 MB	NPU
GPUNet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	0.778 ms	0 - 62 MB	NPU
GPUNet	ONNX	float	Qualcomm® QCS8550 (Proxy)	1.039 ms	0 - 2 MB	NPU
GPUNet	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	0.605 ms	0 - 36 MB	NPU
GPUNet	ONNX	float	Qualcomm® QCS9075	1.353 ms	1 - 3 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	0.355 ms	0 - 48 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® 8 Elite Mobile	0.433 ms	0 - 42 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® X2 Elite	0.384 ms	12 - 12 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® X Elite	0.974 ms	12 - 12 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® X Elite	0.974 ms	12 - 12 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	0.6 ms	0 - 61 MB	NPU
GPUNet	ONNX	w8a16	Qualcomm® QCS6490	98.025 ms	25 - 33 MB	CPU
GPUNet	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	0.825 ms	0 - 16 MB	NPU
GPUNet	ONNX	w8a16	Qualcomm® QCM6690	49.714 ms	29 - 37 MB	CPU
GPUNet	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	38.185 ms	29 - 37 MB	CPU
GPUNet	ONNX	w8a16	Qualcomm® QCS9075	0.98 ms	0 - 3 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	0.433 ms	0 - 42 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	38.185 ms	29 - 37 MB	CPU
GPUNet	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.358 ms	0 - 48 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® 8 Elite Mobile	0.395 ms	0 - 44 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® X2 Elite	0.308 ms	12 - 12 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® X Elite	0.722 ms	12 - 12 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® X Elite	0.722 ms	12 - 12 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.462 ms	0 - 62 MB	NPU
GPUNet	ONNX	w8a8	Qualcomm® QCS6490	16.619 ms	1 - 14 MB	CPU
GPUNet	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	0.616 ms	0 - 22 MB	NPU
GPUNet	ONNX	w8a8	Qualcomm® QCS9075	0.723 ms	0 - 3 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	7.765 ms	7 - 16 MB	CPU
GPUNet	ONNX	w8a8	Qualcomm® QCM6690	10.202 ms	7 - 15 MB	CPU
GPUNet	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.395 ms	0 - 44 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	7.765 ms	7 - 16 MB	CPU
GPUNet	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	0.566 ms	1 - 35 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® 8 Elite Mobile	0.704 ms	0 - 30 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® X2 Elite	0.64 ms	1 - 1 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® X Elite	1.391 ms	1 - 1 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® X Elite	1.391 ms	1 - 1 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.91 ms	0 - 56 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	1.276 ms	1 - 26 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA8775P	1.701 ms	1 - 34 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA8775P	1.701 ms	1 - 34 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA8775P	1.701 ms	1 - 34 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	2.418 ms	0 - 63 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA7255P	4.676 ms	1 - 31 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	0.704 ms	0 - 30 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA8295P	2.265 ms	0 - 31 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® QCS9075	1.572 ms	1 - 3 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	0.434 ms	0 - 44 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	0.525 ms	0 - 45 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® X2 Elite	0.56 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® X Elite	1.243 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® X Elite	1.243 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	0.767 ms	0 - 58 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS6490	3.256 ms	0 - 2 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	1.069 ms	1 - 71 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA8775P	1.274 ms	0 - 44 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA8775P	1.274 ms	0 - 44 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA8775P	1.274 ms	0 - 44 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCM6690	6.476 ms	0 - 166 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS8450 (Proxy)	1.448 ms	0 - 60 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	1.294 ms	0 - 42 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS9075	1.228 ms	0 - 2 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA7255P	2.455 ms	0 - 42 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	0.525 ms	0 - 45 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA8295P	1.649 ms	0 - 39 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	1.294 ms	0 - 42 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.287 ms	0 - 43 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 8 Elite Mobile	0.337 ms	0 - 37 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.347 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® X Elite	0.724 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® X Elite	0.724 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.46 ms	0 - 55 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS6490	1.946 ms	0 - 2 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	0.612 ms	0 - 13 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA8775P	0.797 ms	0 - 40 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA8775P	0.797 ms	0 - 40 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA8775P	0.797 ms	0 - 40 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS9075	0.712 ms	2 - 4 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	0.849 ms	0 - 58 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	0.791 ms	0 - 40 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCM6690	3.458 ms	0 - 43 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.337 ms	0 - 37 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA7255P	1.413 ms	0 - 39 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA8295P	1.11 ms	0 - 36 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	0.791 ms	0 - 40 MB	NPU
GPUNet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	0.559 ms	0 - 47 MB	NPU
GPUNet	TFLITE	float	Snapdragon® 8 Elite Mobile	0.698 ms	0 - 42 MB	NPU
GPUNet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.903 ms	0 - 74 MB	NPU
GPUNet	TFLITE	float	Qualcomm® QCS8550 (Proxy)	1.263 ms	0 - 31 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA8775P	1.754 ms	0 - 46 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA8775P	1.754 ms	0 - 46 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA8775P	1.754 ms	0 - 46 MB	NPU
GPUNet	TFLITE	float	Qualcomm® QCS8450 (Proxy)	2.41 ms	0 - 79 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA7255P	4.739 ms	0 - 43 MB	NPU
GPUNet	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	0.698 ms	0 - 42 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA8295P	2.264 ms	0 - 44 MB	NPU
GPUNet	TFLITE	float	Qualcomm® QCS9075	1.576 ms	0 - 27 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.237 ms	0 - 45 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 8 Elite Mobile	0.274 ms	0 - 39 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.344 ms	0 - 55 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS6490	1.583 ms	0 - 15 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	0.43 ms	0 - 2 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA8775P	0.63 ms	0 - 42 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA8775P	0.63 ms	0 - 42 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA8775P	0.63 ms	0 - 42 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS9075	0.529 ms	0 - 14 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	0.687 ms	0 - 57 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	0.616 ms	0 - 39 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCM6690	3.051 ms	0 - 41 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.274 ms	0 - 39 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA7255P	1.119 ms	0 - 39 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA8295P	0.917 ms	0 - 36 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	0.616 ms	0 - 39 MB	NPU

License

• The license for the original implementation of GPUNet can be found here.

References

• GPUNet: Searching the Deployable Convolution Neural Networks for GPUs
• 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.