MNASNet05: Optimized for Qualcomm Devices

MNASNet05 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 MNASNet05 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 2.45, ONNX Runtime 1.27.1	Download
ONNX	w8a16	Universal	QAIRT 2.45, ONNX Runtime 1.27.1	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a16	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit MNASNet05 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 MNASNet05 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

• Input resolution: 224x224
• Model checkpoint: Imagenet
• Model size (float): 8.45 MB
• Model size (w8a16): 2.79 MB
• Number of parameters: 2.21M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
MNASNet05	ONNX	float	Snapdragon® X2 Elite	0.235 ms	2 - 2 MB	NPU
MNASNet05	ONNX	float	Snapdragon® X Elite	0.484 ms	5 - 5 MB	NPU
MNASNet05	ONNX	float	Snapdragon® 8 Gen 3 Mobile	0.332 ms	0 - 47 MB	NPU
MNASNet05	ONNX	float	Snapdragon® 8 Gen 1 Mobile	0.872 ms	1 - 52 MB	NPU
MNASNet05	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	0.7 ms	0 - 4 MB	NPU
MNASNet05	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.487 ms	0 - 7 MB	NPU
MNASNet05	ONNX	float	Qualcomm® QCS8450	0.872 ms	1 - 52 MB	NPU
MNASNet05	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	0.704 ms	1 - 3 MB	NPU
MNASNet05	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	0.484 ms	5 - 5 MB	NPU
MNASNet05	ONNX	float	Qualcomm® Dragonwing™ Q-8750	0.269 ms	0 - 28 MB	NPU
MNASNet05	ONNX	float	Snapdragon® 8 Elite Mobile	0.269 ms	0 - 28 MB	NPU
MNASNet05	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	0.231 ms	0 - 29 MB	NPU
MNASNet05	ONNX	w8a16	Snapdragon® X2 Elite	0.219 ms	1 - 1 MB	NPU
MNASNet05	ONNX	w8a16	Snapdragon® X Elite	0.515 ms	0 - 0 MB	NPU
MNASNet05	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	0.345 ms	0 - 44 MB	NPU
MNASNet05	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	0.669 ms	0 - 47 MB	NPU
MNASNet05	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	2.163 ms	0 - 3 MB	NPU
MNASNet05	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-8275	0.754 ms	0 - 4 MB	NPU
MNASNet05	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.515 ms	0 - 5 MB	NPU
MNASNet05	ONNX	w8a16	Qualcomm® QCS8450	0.669 ms	0 - 47 MB	NPU
MNASNet05	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	0.652 ms	0 - 3 MB	NPU
MNASNet05	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	0.515 ms	0 - 0 MB	NPU
MNASNet05	ONNX	w8a16	Qualcomm® Dragonwing™ Q-6690	2.239 ms	0 - 145 MB	NPU
MNASNet05	ONNX	w8a16	Qualcomm® Dragonwing™ Q-7790	0.536 ms	0 - 31 MB	NPU
MNASNet05	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	0.261 ms	0 - 30 MB	NPU
MNASNet05	ONNX	w8a16	Snapdragon® 8 Elite Mobile	0.261 ms	0 - 30 MB	NPU
MNASNet05	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	0.219 ms	0 - 33 MB	NPU
MNASNet05	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	0.536 ms	0 - 31 MB	NPU
MNASNet05	QNN_DLC	float	Snapdragon® X2 Elite	0.402 ms	1 - 1 MB	NPU
MNASNet05	QNN_DLC	float	Snapdragon® X Elite	0.915 ms	1 - 1 MB	NPU
MNASNet05	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.509 ms	0 - 45 MB	NPU
MNASNet05	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	1.564 ms	0 - 51 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	0.906 ms	1 - 4 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.788 ms	0 - 3 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® SA8775P	1.103 ms	1 - 30 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® SA8650P	1.103 ms	1 - 30 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® SA8255P	1.103 ms	1 - 30 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® QCS8450	1.564 ms	0 - 51 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	0.988 ms	1 - 3 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	0.915 ms	1 - 1 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	0.385 ms	1 - 32 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® SA7255P	2.305 ms	1 - 28 MB	NPU
MNASNet05	QNN_DLC	float	Qualcomm® SA8295P	1.418 ms	0 - 27 MB	NPU
MNASNet05	QNN_DLC	float	Snapdragon® 8 Elite Mobile	0.385 ms	1 - 32 MB	NPU
MNASNet05	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	0.295 ms	1 - 29 MB	NPU
MNASNet05	QNN_DLC	w8a16	Snapdragon® X2 Elite	0.408 ms	0 - 0 MB	NPU
MNASNet05	QNN_DLC	w8a16	Snapdragon® X Elite	0.907 ms	0 - 0 MB	NPU
MNASNet05	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	0.517 ms	0 - 38 MB	NPU
MNASNet05	QNN_DLC	w8a16	Snapdragon® 8 Gen 1 Mobile	0.956 ms	0 - 45 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS6490	2.747 ms	2 - 4 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	0.768 ms	0 - 3 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.764 ms	0 - 23 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® SA8775P	0.951 ms	0 - 29 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® SA8650P	0.951 ms	0 - 29 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® SA8255P	0.951 ms	0 - 29 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® QCS8450	0.956 ms	0 - 45 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	0.853 ms	0 - 2 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	0.907 ms	0 - 0 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-6690	3.047 ms	0 - 140 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-7790	0.79 ms	0 - 27 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	0.354 ms	0 - 27 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® SA7255P	1.682 ms	0 - 28 MB	NPU
MNASNet05	QNN_DLC	w8a16	Qualcomm® SA8295P	1.225 ms	0 - 24 MB	NPU
MNASNet05	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	0.354 ms	0 - 27 MB	NPU
MNASNet05	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	0.305 ms	0 - 27 MB	NPU
MNASNet05	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	0.79 ms	0 - 27 MB	NPU
MNASNet05	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.518 ms	0 - 46 MB	NPU
MNASNet05	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	1.57 ms	0 - 52 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	0.915 ms	0 - 9 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.797 ms	0 - 1 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® SA8775P	1.116 ms	0 - 31 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® SA8650P	1.116 ms	0 - 31 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® SA8255P	1.116 ms	0 - 31 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® QCS8450	1.57 ms	0 - 52 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	0.98 ms	0 - 8 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	0.382 ms	0 - 29 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® SA7255P	2.325 ms	0 - 28 MB	NPU
MNASNet05	TFLITE	float	Qualcomm® SA8295P	1.481 ms	0 - 28 MB	NPU
MNASNet05	TFLITE	float	Snapdragon® 8 Elite Mobile	0.382 ms	0 - 29 MB	NPU
MNASNet05	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	0.295 ms	0 - 31 MB	NPU

License

• The license for the original implementation of MNASNet05 can be found here.

References

• MnasNet: Platform-Aware Neural Architecture Search for Mobile
• 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.