Mobile-VIT: Optimized for Qualcomm Devices

MobileVit 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 Mobile-VIT 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
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 Mobile-VIT 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 Mobile-VIT 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: 5.57M
Model size (float): 21.4 MB
Model size (w8a16): 6.56 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Mobile-VIT	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	2.021 ms	1 - 94 MB	NPU
Mobile-VIT	ONNX	float	Snapdragon® 8 Elite Mobile	2.456 ms	0 - 78 MB	NPU
Mobile-VIT	ONNX	float	Snapdragon® X2 Elite	2.268 ms	12 - 12 MB	NPU
Mobile-VIT	ONNX	float	Snapdragon® X Elite	4.782 ms	12 - 12 MB	NPU
Mobile-VIT	ONNX	float	Snapdragon® X Elite	4.782 ms	12 - 12 MB	NPU
Mobile-VIT	ONNX	float	Snapdragon® 8 Gen 3 Mobile	3.029 ms	0 - 99 MB	NPU
Mobile-VIT	ONNX	float	Qualcomm® QCS8550 (Proxy)	4.385 ms	0 - 17 MB	NPU
Mobile-VIT	ONNX	float	Qualcomm® QCS9075	5.204 ms	1 - 4 MB	NPU
Mobile-VIT	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	2.456 ms	0 - 78 MB	NPU
Mobile-VIT	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	2.324 ms	0 - 103 MB	NPU
Mobile-VIT	ONNX	w8a16	Snapdragon® 8 Elite Mobile	2.743 ms	0 - 110 MB	NPU
Mobile-VIT	ONNX	w8a16	Snapdragon® X2 Elite	2.473 ms	6 - 6 MB	NPU
Mobile-VIT	ONNX	w8a16	Snapdragon® X Elite	5.907 ms	6 - 6 MB	NPU
Mobile-VIT	ONNX	w8a16	Snapdragon® X Elite	5.907 ms	6 - 6 MB	NPU
Mobile-VIT	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	3.644 ms	0 - 121 MB	NPU
Mobile-VIT	ONNX	w8a16	Qualcomm® QCS6490	320.267 ms	63 - 67 MB	CPU
Mobile-VIT	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	5.481 ms	0 - 10 MB	NPU
Mobile-VIT	ONNX	w8a16	Qualcomm® QCS9075	5.815 ms	0 - 3 MB	NPU
Mobile-VIT	ONNX	w8a16	Qualcomm® QCM6690	139.206 ms	63 - 75 MB	CPU
Mobile-VIT	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	2.743 ms	0 - 110 MB	NPU
Mobile-VIT	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	121.198 ms	63 - 74 MB	CPU
Mobile-VIT	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	121.198 ms	63 - 74 MB	CPU
Mobile-VIT	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	1.973 ms	1 - 61 MB	NPU
Mobile-VIT	QNN_DLC	float	Snapdragon® 8 Elite Mobile	2.386 ms	0 - 59 MB	NPU
Mobile-VIT	QNN_DLC	float	Snapdragon® X2 Elite	2.494 ms	1 - 1 MB	NPU
Mobile-VIT	QNN_DLC	float	Snapdragon® X Elite	4.615 ms	1 - 1 MB	NPU
Mobile-VIT	QNN_DLC	float	Snapdragon® X Elite	4.615 ms	1 - 1 MB	NPU
Mobile-VIT	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	2.938 ms	0 - 88 MB	NPU
Mobile-VIT	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	11.191 ms	1 - 57 MB	NPU
Mobile-VIT	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	4.259 ms	1 - 3 MB	NPU
Mobile-VIT	QNN_DLC	float	Qualcomm® SA8775P	5.023 ms	1 - 66 MB	NPU
Mobile-VIT	QNN_DLC	float	Qualcomm® SA8775P	5.023 ms	1 - 66 MB	NPU
Mobile-VIT	QNN_DLC	float	Qualcomm® SA8775P	5.023 ms	1 - 66 MB	NPU
Mobile-VIT	QNN_DLC	float	Qualcomm® QCS9075	4.963 ms	1 - 3 MB	NPU
Mobile-VIT	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	6.428 ms	0 - 88 MB	NPU
Mobile-VIT	QNN_DLC	float	Qualcomm® SA7255P	11.191 ms	1 - 57 MB	NPU
Mobile-VIT	QNN_DLC	float	Qualcomm® SA8295P	7.35 ms	1 - 61 MB	NPU
Mobile-VIT	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	2.386 ms	0 - 59 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	1.785 ms	0 - 76 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	2.187 ms	0 - 73 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Snapdragon® X2 Elite	2.19 ms	0 - 0 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Snapdragon® X Elite	4.902 ms	0 - 0 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Snapdragon® X Elite	4.902 ms	0 - 0 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	2.973 ms	0 - 93 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Qualcomm® QCS8275 (Proxy)	8.055 ms	0 - 71 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	4.467 ms	0 - 3 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Qualcomm® SA8775P	4.918 ms	0 - 70 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Qualcomm® SA8775P	4.918 ms	0 - 70 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Qualcomm® SA8775P	4.918 ms	0 - 70 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Qualcomm® QCS9075	4.863 ms	2 - 4 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Qualcomm® QCM6690	18.915 ms	0 - 224 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Qualcomm® SA7255P	8.055 ms	0 - 71 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	2.187 ms	0 - 73 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	4.858 ms	0 - 217 MB	NPU
Mobile-VIT	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	4.858 ms	0 - 217 MB	NPU
Mobile-VIT	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	1.974 ms	0 - 77 MB	NPU
Mobile-VIT	TFLITE	float	Snapdragon® 8 Elite Mobile	2.386 ms	0 - 77 MB	NPU
Mobile-VIT	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	2.946 ms	0 - 97 MB	NPU
Mobile-VIT	TFLITE	float	Qualcomm® QCS8275 (Proxy)	11.205 ms	0 - 78 MB	NPU
Mobile-VIT	TFLITE	float	Qualcomm® QCS8550 (Proxy)	4.215 ms	0 - 2 MB	NPU
Mobile-VIT	TFLITE	float	Qualcomm® SA8775P	5.063 ms	0 - 80 MB	NPU
Mobile-VIT	TFLITE	float	Qualcomm® SA8775P	5.063 ms	0 - 80 MB	NPU
Mobile-VIT	TFLITE	float	Qualcomm® SA8775P	5.063 ms	0 - 80 MB	NPU
Mobile-VIT	TFLITE	float	Qualcomm® QCS9075	4.976 ms	0 - 15 MB	NPU
Mobile-VIT	TFLITE	float	Qualcomm® QCS8450 (Proxy)	6.424 ms	0 - 98 MB	NPU
Mobile-VIT	TFLITE	float	Qualcomm® SA7255P	11.205 ms	0 - 78 MB	NPU
Mobile-VIT	TFLITE	float	Qualcomm® SA8295P	7.407 ms	0 - 64 MB	NPU
Mobile-VIT	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	2.386 ms	0 - 77 MB	NPU

License

The license for the original implementation of Mobile-VIT 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/Mobile-VIT

MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer

Paper • 2110.02178 • Published Oct 5, 2021 • 3