Swin-Tiny: Optimized for Qualcomm Devices

SwinTiny 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-Tiny 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 1.25.0	Download
ONNX	w8a16	Universal	QAIRT 2.45, ONNX Runtime 1.25.0	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 Swin-Tiny 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-Tiny 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: 28.8M
• Model size (float): 110 MB
• Model size (w8a16): 29.9 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Swin-Tiny	ONNX	float	Snapdragon® X2 Elite	3.972 ms	2 - 2 MB	NPU
Swin-Tiny	ONNX	float	Snapdragon® X Elite	9.65 ms	58 - 58 MB	NPU
Swin-Tiny	ONNX	float	Snapdragon® 8 Gen 3 Mobile	5.995 ms	0 - 252 MB	NPU
Swin-Tiny	ONNX	float	Snapdragon® 8 Gen 1 Mobile	14.245 ms	1 - 247 MB	NPU
Swin-Tiny	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	9.112 ms	0 - 66 MB	NPU
Swin-Tiny	ONNX	float	Qualcomm® QCS8450	14.245 ms	1 - 247 MB	NPU
Swin-Tiny	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	10.187 ms	1 - 4 MB	NPU
Swin-Tiny	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	3.775 ms	0 - 433 MB	NPU
Swin-Tiny	ONNX	float	Snapdragon® 8 Elite Mobile	4.578 ms	0 - 434 MB	NPU
Swin-Tiny	ONNX	float	Qualcomm® Dragonwing™ Q-8750	4.578 ms	0 - 434 MB	NPU
Swin-Tiny	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	9.65 ms	58 - 58 MB	NPU
Swin-Tiny	ONNX	w8a16	Snapdragon® X2 Elite	3.476 ms	1 - 1 MB	NPU
Swin-Tiny	ONNX	w8a16	Snapdragon® X Elite	8.287 ms	31 - 31 MB	NPU
Swin-Tiny	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	5.163 ms	0 - 333 MB	NPU
Swin-Tiny	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	7.81 ms	0 - 36 MB	NPU
Swin-Tiny	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	8.116 ms	0 - 3 MB	NPU
Swin-Tiny	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	8.241 ms	0 - 237 MB	NPU
Swin-Tiny	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	3.273 ms	0 - 275 MB	NPU
Swin-Tiny	ONNX	w8a16	Qualcomm® Dragonwing™ Q-6690	34.435 ms	0 - 394 MB	NPU
Swin-Tiny	ONNX	w8a16	Snapdragon® 8 Elite Mobile	3.973 ms	0 - 267 MB	NPU
Swin-Tiny	ONNX	w8a16	Qualcomm® Dragonwing™ Q-7790	8.241 ms	0 - 237 MB	NPU
Swin-Tiny	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	3.973 ms	0 - 267 MB	NPU
Swin-Tiny	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	8.287 ms	31 - 31 MB	NPU
Swin-Tiny	QNN_DLC	float	Snapdragon® X2 Elite	4.285 ms	1 - 1 MB	NPU
Swin-Tiny	QNN_DLC	float	Snapdragon® X Elite	9.862 ms	1 - 1 MB	NPU
Swin-Tiny	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	5.945 ms	0 - 246 MB	NPU
Swin-Tiny	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	14.509 ms	0 - 240 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® QCS8275	21.257 ms	1 - 176 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	9.12 ms	1 - 2 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® SA8775P	10.453 ms	1 - 436 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® SA8650P	10.453 ms	1 - 436 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® SA8255P	10.453 ms	1 - 436 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® QCS8450	14.509 ms	0 - 240 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	10.243 ms	1 - 3 MB	NPU
Swin-Tiny	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	3.773 ms	1 - 414 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® SA7255P	21.257 ms	1 - 176 MB	NPU
Swin-Tiny	QNN_DLC	float	Snapdragon® 8 Elite Mobile	4.493 ms	0 - 414 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® SA8295P	13.817 ms	0 - 432 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	4.493 ms	0 - 414 MB	NPU
Swin-Tiny	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	9.862 ms	1 - 1 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Snapdragon® X2 Elite	4.416 ms	0 - 0 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Snapdragon® X Elite	11.025 ms	0 - 0 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	6.652 ms	0 - 508 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® QCS8275	17.763 ms	0 - 436 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	10.171 ms	0 - 9 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® SA8775P	10.663 ms	0 - 444 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® SA8650P	10.663 ms	0 - 444 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® SA8255P	10.663 ms	0 - 444 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	10.445 ms	2 - 4 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	10.633 ms	0 - 437 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	3.892 ms	0 - 439 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-6690	42.42 ms	0 - 430 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® SA7255P	17.763 ms	0 - 436 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	4.843 ms	0 - 428 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-7790	10.633 ms	0 - 437 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	4.843 ms	0 - 428 MB	NPU
Swin-Tiny	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	11.025 ms	0 - 0 MB	NPU
Swin-Tiny	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	6.008 ms	0 - 258 MB	NPU
Swin-Tiny	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	14.318 ms	0 - 256 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® QCS8275	21.269 ms	0 - 200 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	9.243 ms	0 - 4 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® SA8775P	10.403 ms	0 - 201 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® SA8650P	10.403 ms	0 - 201 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® SA8255P	10.403 ms	0 - 201 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® QCS8450	14.318 ms	0 - 256 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	10.161 ms	0 - 60 MB	NPU
Swin-Tiny	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	3.824 ms	0 - 214 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® SA7255P	21.269 ms	0 - 200 MB	NPU
Swin-Tiny	TFLITE	float	Snapdragon® 8 Elite Mobile	4.64 ms	0 - 205 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® SA8295P	13.889 ms	0 - 199 MB	NPU
Swin-Tiny	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	4.64 ms	0 - 205 MB	NPU

License

• The license for the original implementation of Swin-Tiny 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.