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 1.25.0	Download
ONNX	w8a16	Universal	QAIRT 2.42, ONNX Runtime 1.25.0	Download
ONNX	w8a8	Universal	QAIRT 2.42, ONNX Runtime 1.25.0	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	0.505 ms	1 - 36 MB	NPU
GPUNet	ONNX	float	Snapdragon® X2 Elite	0.463 ms	212 - 212 MB	NPU
GPUNet	ONNX	float	Snapdragon® X Elite	1.006 ms	149 - 149 MB	NPU
GPUNet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	0.775 ms	0 - 63 MB	NPU
GPUNet	ONNX	float	Qualcomm® QCS8550 (Proxy)	1.049 ms	0 - 46 MB	NPU
GPUNet	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	0.614 ms	0 - 36 MB	NPU
GPUNet	ONNX	float	Qualcomm® QCS9075	1.368 ms	1 - 46 MB	NPU
GPUNet	ONNX	float	Qualcomm® QCS8750	0.614 ms	0 - 36 MB	NPU
GPUNet	ONNX	float	Qualcomm® QCS7181	1.006 ms	149 - 149 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	0.352 ms	0 - 50 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® X Elite	0.85 ms	149 - 149 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	0.589 ms	0 - 69 MB	NPU
GPUNet	ONNX	w8a16	Qualcomm® QCS6490	97.97 ms	24 - 33 MB	CPU
GPUNet	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	0.865 ms	0 - 26 MB	NPU
GPUNet	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	37.9 ms	30 - 37 MB	CPU
GPUNet	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	0.431 ms	0 - 48 MB	NPU
GPUNet	ONNX	w8a16	Qualcomm® QCM6690	49.44 ms	30 - 38 MB	CPU
GPUNet	ONNX	w8a16	Qualcomm® QCS9075	0.983 ms	0 - 45 MB	NPU
GPUNet	ONNX	w8a16	Qualcomm® QCS7790	37.9 ms	30 - 37 MB	CPU
GPUNet	ONNX	w8a16	Qualcomm® QCS8750	0.431 ms	0 - 48 MB	NPU
GPUNet	ONNX	w8a16	Qualcomm® QCS7181	0.85 ms	149 - 149 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.352 ms	0 - 50 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® X2 Elite	0.31 ms	212 - 212 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® X Elite	0.604 ms	149 - 149 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.465 ms	0 - 68 MB	NPU
GPUNet	ONNX	w8a8	Qualcomm® QCS6490	16.454 ms	1 - 14 MB	CPU
GPUNet	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	0.619 ms	0 - 47 MB	NPU
GPUNet	ONNX	w8a8	Qualcomm® QCS9075	0.725 ms	0 - 45 MB	NPU
GPUNet	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	7.569 ms	7 - 15 MB	CPU
GPUNet	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.391 ms	0 - 43 MB	NPU
GPUNet	ONNX	w8a8	Qualcomm® QCM6690	10.0 ms	6 - 15 MB	CPU
GPUNet	ONNX	w8a8	Qualcomm® QCS7790	7.569 ms	7 - 15 MB	CPU
GPUNet	ONNX	w8a8	Qualcomm® QCS8750	0.391 ms	0 - 43 MB	NPU
GPUNet	ONNX	w8a8	Qualcomm® QCS7181	0.604 ms	149 - 149 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	0.567 ms	1 - 35 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® X2 Elite	0.636 ms	1 - 1 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® X Elite	1.386 ms	1 - 1 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.916 ms	0 - 57 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® QCS8275	4.679 ms	1 - 31 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	1.27 ms	1 - 2 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA8775P	1.706 ms	1 - 33 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA8650P	1.706 ms	1 - 33 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA8255P	1.706 ms	1 - 33 MB	NPU
GPUNet	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	0.702 ms	1 - 36 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	2.413 ms	0 - 60 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA7255P	4.679 ms	1 - 31 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® QCS9075	1.573 ms	3 - 5 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® SA8295P	2.285 ms	0 - 32 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® QCS8750	0.702 ms	1 - 36 MB	NPU
GPUNet	QNN_DLC	float	Qualcomm® QCS7181	1.386 ms	1 - 1 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	0.43 ms	0 - 46 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® X2 Elite	0.563 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® X Elite	1.228 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	0.772 ms	0 - 59 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS6490	3.243 ms	2 - 4 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS8275	2.482 ms	0 - 43 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	1.068 ms	0 - 2 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA8775P	1.293 ms	0 - 45 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA8650P	1.293 ms	0 - 45 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA8255P	1.293 ms	0 - 45 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	1.292 ms	0 - 43 MB	NPU
GPUNet	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	0.528 ms	0 - 42 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCM6690	6.561 ms	0 - 168 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA7255P	2.482 ms	0 - 43 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® SA8295P	1.651 ms	0 - 40 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS9075	1.239 ms	2 - 4 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS8450 (Proxy)	1.461 ms	0 - 62 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS7790	1.292 ms	0 - 43 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS8750	0.528 ms	0 - 42 MB	NPU
GPUNet	QNN_DLC	w8a16	Qualcomm® QCS7181	1.228 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.296 ms	0 - 45 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.35 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® X Elite	0.701 ms	0 - 0 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.466 ms	0 - 57 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS6490	2.045 ms	2 - 4 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS8275	1.412 ms	0 - 40 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	0.614 ms	0 - 2 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA8775P	0.816 ms	0 - 42 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA8650P	0.816 ms	0 - 42 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA8255P	0.816 ms	0 - 42 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS9075	0.687 ms	0 - 2 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	0.857 ms	0 - 59 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	0.784 ms	0 - 40 MB	NPU
GPUNet	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.346 ms	0 - 39 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCM6690	3.565 ms	0 - 44 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA8295P	1.106 ms	0 - 38 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® SA7255P	1.412 ms	0 - 40 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS7790	0.784 ms	0 - 40 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS8750	0.346 ms	0 - 39 MB	NPU
GPUNet	QNN_DLC	w8a8	Qualcomm® QCS7181	0.701 ms	0 - 0 MB	NPU
GPUNet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	0.557 ms	0 - 47 MB	NPU
GPUNet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.906 ms	0 - 74 MB	NPU
GPUNet	TFLITE	float	Qualcomm® QCS8275	4.686 ms	0 - 43 MB	NPU
GPUNet	TFLITE	float	Qualcomm® QCS8550 (Proxy)	1.263 ms	0 - 16 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA8775P	1.748 ms	0 - 47 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA8650P	1.748 ms	0 - 47 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA8255P	1.748 ms	0 - 47 MB	NPU
GPUNet	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	0.699 ms	0 - 47 MB	NPU
GPUNet	TFLITE	float	Qualcomm® QCS8450 (Proxy)	2.407 ms	0 - 78 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA7255P	4.686 ms	0 - 43 MB	NPU
GPUNet	TFLITE	float	Qualcomm® QCS9075	1.582 ms	0 - 27 MB	NPU
GPUNet	TFLITE	float	Qualcomm® SA8295P	2.273 ms	0 - 43 MB	NPU
GPUNet	TFLITE	float	Qualcomm® QCS8750	0.699 ms	0 - 47 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.236 ms	0 - 45 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.345 ms	0 - 56 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS6490	1.642 ms	0 - 15 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS8275	1.113 ms	0 - 39 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	0.428 ms	0 - 3 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA8775P	0.629 ms	0 - 42 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA8650P	0.629 ms	0 - 42 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA8255P	0.629 ms	0 - 42 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS9075	0.509 ms	0 - 14 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	0.684 ms	0 - 57 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	0.613 ms	0 - 39 MB	NPU
GPUNet	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.279 ms	0 - 42 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCM6690	2.961 ms	0 - 40 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA8295P	0.91 ms	0 - 37 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® SA7255P	1.113 ms	0 - 39 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS7790	0.613 ms	0 - 39 MB	NPU
GPUNet	TFLITE	w8a8	Qualcomm® QCS8750	0.279 ms	0 - 42 MB	NPU

License

The license for the original implementation of GPUNet 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.

Downloads last month: -; Downloads are not tracked for this model. How to track

Paper for qualcomm/GPUNet

GPUNet: Searching the Deployable Convolution Neural Networks for GPUs

Paper • 2205.00841 • Published Apr 26, 2022