Swin-Small: Optimized for Qualcomm Devices

SwinSmall 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 Swin-Small 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 2.50, ONNX Runtime 1.27.1	Download
ONNX	w8a16	Universal	QAIRT 2.50, ONNX Runtime 1.27.1	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 Swin-Small 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 Swin-Small 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): 193 MB
• Model size (w8a16): 52.5 MB
• Number of parameters: 50.4M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Swin-Small	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.249 ms	1 - 320 MB	NPU
Swin-Small	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	7.642 ms	1 - 307 MB	NPU
Swin-Small	ONNX	float	Snapdragon® X2 Elite	6.437 ms	2 - 2 MB	NPU
Swin-Small	ONNX	float	Snapdragon® X Elite	15.714 ms	101 - 101 MB	NPU
Swin-Small	ONNX	float	Snapdragon® 8 Gen 3 Mobile	9.971 ms	0 - 438 MB	NPU
Swin-Small	ONNX	float	Snapdragon® 8 Gen 1 Mobile	23.356 ms	0 - 425 MB	NPU
Swin-Small	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	15.275 ms	0 - 5 MB	NPU
Swin-Small	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	15.184 ms	0 - 114 MB	NPU
Swin-Small	ONNX	float	Qualcomm® QCS8450	23.356 ms	0 - 425 MB	NPU
Swin-Small	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	16.724 ms	1 - 5 MB	NPU
Swin-Small	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	15.714 ms	101 - 101 MB	NPU
Swin-Small	ONNX	float	Qualcomm® Dragonwing™ Q-8750	7.642 ms	1 - 307 MB	NPU
Swin-Small	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	5.153 ms	0 - 340 MB	NPU
Swin-Small	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	6.605 ms	0 - 330 MB	NPU
Swin-Small	ONNX	w8a16	Snapdragon® X2 Elite	5.364 ms	1 - 1 MB	NPU
Swin-Small	ONNX	w8a16	Snapdragon® X Elite	13.505 ms	54 - 54 MB	NPU
Swin-Small	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	8.856 ms	0 - 444 MB	NPU
Swin-Small	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	16.454 ms	0 - 441 MB	NPU
Swin-Small	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	26.821 ms	0 - 4 MB	NPU
Swin-Small	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-8275	11.558 ms	0 - 5 MB	NPU
Swin-Small	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	13.067 ms	0 - 59 MB	NPU
Swin-Small	ONNX	w8a16	Qualcomm® QCS8450	16.454 ms	0 - 441 MB	NPU
Swin-Small	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	13.688 ms	0 - 4 MB	NPU
Swin-Small	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	13.505 ms	54 - 54 MB	NPU
Swin-Small	ONNX	w8a16	Qualcomm® Dragonwing™ Q-6690	66.905 ms	0 - 535 MB	NPU
Swin-Small	ONNX	w8a16	Qualcomm® Dragonwing™ Q-7790	14.201 ms	0 - 374 MB	NPU
Swin-Small	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	6.605 ms	0 - 330 MB	NPU
Swin-Small	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	14.201 ms	0 - 374 MB	NPU
Swin-Small	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.098 ms	0 - 286 MB	NPU
Swin-Small	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	7.447 ms	1 - 270 MB	NPU
Swin-Small	QNN_DLC	float	Snapdragon® X2 Elite	6.785 ms	1 - 1 MB	NPU
Swin-Small	QNN_DLC	float	Snapdragon® X Elite	15.78 ms	1 - 1 MB	NPU
Swin-Small	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	9.87 ms	0 - 402 MB	NPU
Swin-Small	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	23.608 ms	0 - 393 MB	NPU
Swin-Small	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	15.206 ms	1 - 4 MB	NPU
Swin-Small	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	14.931 ms	1 - 3 MB	NPU
Swin-Small	QNN_DLC	float	Qualcomm® SA8650P	17.421 ms	1 - 593 MB	NPU
Swin-Small	QNN_DLC	float	Qualcomm® SA8255P	17.421 ms	1 - 593 MB	NPU
Swin-Small	QNN_DLC	float	Qualcomm® QCS8450	23.608 ms	0 - 393 MB	NPU
Swin-Small	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	17.142 ms	3 - 5 MB	NPU
Swin-Small	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	15.78 ms	1 - 1 MB	NPU
Swin-Small	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	7.447 ms	1 - 270 MB	NPU
Swin-Small	QNN_DLC	float	Qualcomm® SA8295P	22.117 ms	1 - 263 MB	NPU
Swin-Small	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.033 ms	0 - 316 MB	NPU
Swin-Small	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	7.854 ms	0 - 305 MB	NPU
Swin-Small	QNN_DLC	w8a16	Snapdragon® X2 Elite	6.569 ms	0 - 0 MB	NPU
Swin-Small	QNN_DLC	w8a16	Snapdragon® X Elite	17.107 ms	0 - 0 MB	NPU
Swin-Small	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	10.943 ms	0 - 408 MB	NPU
Swin-Small	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	14.468 ms	0 - 4 MB	NPU
Swin-Small	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	16.13 ms	0 - 3 MB	NPU
Swin-Small	QNN_DLC	w8a16	Qualcomm® SA8650P	16.763 ms	0 - 638 MB	NPU
Swin-Small	QNN_DLC	w8a16	Qualcomm® SA8255P	16.763 ms	0 - 638 MB	NPU
Swin-Small	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	17.062 ms	0 - 3 MB	NPU
Swin-Small	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	17.107 ms	0 - 0 MB	NPU
Swin-Small	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-6690	75.035 ms	0 - 512 MB	NPU
Swin-Small	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-7790	17.104 ms	0 - 352 MB	NPU
Swin-Small	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	7.854 ms	0 - 305 MB	NPU
Swin-Small	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	17.104 ms	0 - 352 MB	NPU
Swin-Small	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	6.522 ms	0 - 313 MB	NPU
Swin-Small	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	7.764 ms	0 - 290 MB	NPU
Swin-Small	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	9.993 ms	0 - 419 MB	NPU
Swin-Small	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	23.439 ms	0 - 408 MB	NPU
Swin-Small	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	15.275 ms	0 - 105 MB	NPU
Swin-Small	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	15.191 ms	0 - 5 MB	NPU
Swin-Small	TFLITE	float	Qualcomm® SA8650P	17.46 ms	0 - 306 MB	NPU
Swin-Small	TFLITE	float	Qualcomm® SA8255P	17.46 ms	0 - 306 MB	NPU
Swin-Small	TFLITE	float	Qualcomm® QCS8450	23.439 ms	0 - 408 MB	NPU
Swin-Small	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	17.434 ms	0 - 104 MB	NPU
Swin-Small	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	7.764 ms	0 - 290 MB	NPU
Swin-Small	TFLITE	float	Qualcomm® SA8295P	22.528 ms	0 - 284 MB	NPU

License

• The license for the original implementation of Swin-Small can be found here.

References

• Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
• 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.