SwinV2-Base: Optimized for Qualcomm Devices

SwinV2Base 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 SwinV2-Base 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
QNN_DLC	float	Universal	QAIRT 2.50	Download
QNN_DLC	w8a16	Universal	QAIRT 2.50	Download
TFLITE	float	Universal	QAIRT 2.50	Download

For more device-specific assets and performance metrics, visit SwinV2-Base 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 SwinV2-Base on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

• Input resolution: 256x256
• Model checkpoint: Imagenet
• Model size (float): 339 MB
• Model size (w8a16): 90.2 MB
• Number of parameters: 88.8M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
SwinV2-Base	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	11.064 ms	0 - 433 MB	NPU
SwinV2-Base	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	14.905 ms	0 - 408 MB	NPU
SwinV2-Base	QNN_DLC	float	Snapdragon® X2 Elite	11.673 ms	1 - 1 MB	NPU
SwinV2-Base	QNN_DLC	float	Snapdragon® X Elite	30.282 ms	1 - 1 MB	NPU
SwinV2-Base	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	19.947 ms	0 - 569 MB	NPU
SwinV2-Base	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	40.273 ms	1 - 559 MB	NPU
SwinV2-Base	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	30.139 ms	1 - 4 MB	NPU
SwinV2-Base	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	28.991 ms	1 - 3 MB	NPU
SwinV2-Base	QNN_DLC	float	Qualcomm® QCS8450	40.273 ms	1 - 559 MB	NPU
SwinV2-Base	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	32.735 ms	3 - 5 MB	NPU
SwinV2-Base	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	30.282 ms	1 - 1 MB	NPU
SwinV2-Base	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	14.905 ms	0 - 408 MB	NPU
SwinV2-Base	QNN_DLC	float	Qualcomm® SA8295P	39.817 ms	1 - 394 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	11.005 ms	0 - 533 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	14.639 ms	0 - 466 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Snapdragon® X2 Elite	11.494 ms	0 - 0 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Snapdragon® X Elite	30.909 ms	1 - 1 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	20.409 ms	0 - 599 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	26.381 ms	0 - 3 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	29.835 ms	0 - 369 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Qualcomm® SA8650P	30.025 ms	0 - 869 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Qualcomm® SA8255P	30.025 ms	0 - 869 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	29.898 ms	0 - 3 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	30.909 ms	1 - 1 MB	NPU
SwinV2-Base	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	14.639 ms	0 - 466 MB	NPU
SwinV2-Base	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	11.217 ms	0 - 523 MB	NPU
SwinV2-Base	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	14.593 ms	0 - 848 MB	NPU
SwinV2-Base	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	20.199 ms	0 - 583 MB	NPU
SwinV2-Base	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	39.985 ms	0 - 608 MB	NPU
SwinV2-Base	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	29.195 ms	0 - 181 MB	NPU
SwinV2-Base	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	29.365 ms	0 - 4 MB	NPU
SwinV2-Base	TFLITE	float	Qualcomm® SA8650P	272.16 ms	3 - 27 MB	CPU
SwinV2-Base	TFLITE	float	Qualcomm® SA8255P	272.16 ms	3 - 27 MB	CPU
SwinV2-Base	TFLITE	float	Qualcomm® QCS8450	39.985 ms	0 - 608 MB	NPU
SwinV2-Base	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	31.782 ms	0 - 181 MB	NPU
SwinV2-Base	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	14.593 ms	0 - 848 MB	NPU
SwinV2-Base	TFLITE	float	Qualcomm® SA8295P	39.054 ms	0 - 414 MB	NPU

License

• The license for the original implementation of SwinV2-Base can be found here.

References

• Swin Transformer V2: Scaling Up Capacity and Resolution
• 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.