Swin-Base: Optimized for Qualcomm Devices

SwinBase 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-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
ONNX float Universal QAIRT 2.51, ONNX Runtime 1.30.0 Download
ONNX w8a16 Universal QAIRT 2.51, ONNX Runtime 1.30.0 Download
QNN_DLC float Universal QAIRT 2.51 Download
QNN_DLC w8a16 Universal QAIRT 2.51 Download
TFLITE float Universal QAIRT 2.51 Download

For more device-specific assets and performance metrics, visit Swin-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 Swin-Base 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): 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
Swin-Base ONNX float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 7.389 ms 0 - 389 MB NPU
Swin-Base ONNX float Snapdragon® 8 Elite For Galaxy Mobile 9.157 ms 0 - 372 MB NPU
Swin-Base ONNX float Snapdragon® X2 Elite 7.632 ms 2 - 2 MB NPU
Swin-Base ONNX float Snapdragon® X Elite 18.334 ms 176 - 176 MB NPU
Swin-Base ONNX float Snapdragon® 8 Gen 3 Mobile 12.56 ms 0 - 528 MB NPU
Swin-Base ONNX float Snapdragon® 8 Gen 1 Mobile 27.458 ms 1 - 521 MB NPU
Swin-Base ONNX float Qualcomm® Dragonwing™ IQ-8275 20.31 ms 0 - 5 MB NPU
Swin-Base ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 17.967 ms 0 - 194 MB NPU
Swin-Base ONNX float Qualcomm® QCS8450 27.458 ms 1 - 521 MB NPU
Swin-Base ONNX float Qualcomm® Dragonwing™ IQ-9075 21.321 ms 0 - 4 MB NPU
Swin-Base ONNX float Qualcomm® Dragonwing™ IQ-X7181 18.334 ms 176 - 176 MB NPU
Swin-Base ONNX float Qualcomm® Dragonwing™ Q-8750 9.157 ms 0 - 372 MB NPU
Swin-Base ONNX float Snapdragon® 8 Elite Extreme Gen 6 Mobile 6.754 ms 1 - 381 MB NPU
Swin-Base ONNX float Snapdragon® 8 Elite Gen 6 Mobile 8.523 ms 0 - 377 MB NPU
Swin-Base ONNX w8a16 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 5.971 ms 0 - 413 MB NPU
Swin-Base ONNX w8a16 Snapdragon® 8 Elite For Galaxy Mobile 7.756 ms 0 - 401 MB NPU
Swin-Base ONNX w8a16 Snapdragon® X2 Elite 6.272 ms 1 - 1 MB NPU
Swin-Base ONNX w8a16 Snapdragon® X Elite 15.931 ms 92 - 92 MB NPU
Swin-Base ONNX w8a16 Snapdragon® 8 Gen 3 Mobile 10.723 ms 0 - 538 MB NPU
Swin-Base ONNX w8a16 Snapdragon® 8 Gen 1 Mobile 19.8 ms 0 - 540 MB NPU
Swin-Base ONNX w8a16 Qualcomm® Dragonwing™ QCS6490 37.938 ms 0 - 4 MB NPU
Swin-Base ONNX w8a16 Qualcomm® Dragonwing™ IQ-8275 13.901 ms 0 - 4 MB NPU
Swin-Base ONNX w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 15.295 ms 0 - 99 MB NPU
Swin-Base ONNX w8a16 Qualcomm® QCS8450 19.8 ms 0 - 540 MB NPU
Swin-Base ONNX w8a16 Qualcomm® Dragonwing™ IQ-9075 15.167 ms 0 - 4 MB NPU
Swin-Base ONNX w8a16 Qualcomm® Dragonwing™ IQ-X7181 15.931 ms 92 - 92 MB NPU
Swin-Base ONNX w8a16 Qualcomm® Dragonwing™ Q-6690 86.493 ms 0 - 678 MB NPU
Swin-Base ONNX w8a16 Qualcomm® Dragonwing™ Q-7790 17.241 ms 0 - 610 MB NPU
Swin-Base ONNX w8a16 Qualcomm® Dragonwing™ Q-8750 7.756 ms 0 - 401 MB NPU
Swin-Base ONNX w8a16 Snapdragon® 7 Gen 4 Mobile 17.241 ms 0 - 610 MB NPU
Swin-Base ONNX w8a16 Snapdragon® 8 Elite Extreme Gen 6 Mobile 5.071 ms 0 - 415 MB NPU
Swin-Base ONNX w8a16 Snapdragon® 8 Elite Gen 6 Mobile 6.871 ms 0 - 417 MB NPU
Swin-Base QNN_DLC float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 6.934 ms 0 - 367 MB NPU
Swin-Base QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 8.717 ms 1 - 351 MB NPU
Swin-Base QNN_DLC float Snapdragon® X2 Elite 7.646 ms 1 - 1 MB NPU
Swin-Base QNN_DLC float Snapdragon® X Elite 17.85 ms 1 - 1 MB NPU
Swin-Base QNN_DLC float Snapdragon® 8 Gen 3 Mobile 12.199 ms 1 - 507 MB NPU
Swin-Base QNN_DLC float Snapdragon® 8 Gen 1 Mobile 26.989 ms 0 - 496 MB NPU
Swin-Base QNN_DLC float Qualcomm® Dragonwing™ IQ-8275 19.522 ms 1 - 4 MB NPU
Swin-Base QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 17.163 ms 1 - 3 MB NPU
Swin-Base QNN_DLC float Qualcomm® QCS8450 26.989 ms 0 - 496 MB NPU
Swin-Base QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 20.11 ms 1 - 4 MB NPU
Swin-Base QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 17.85 ms 1 - 1 MB NPU
Swin-Base QNN_DLC float Qualcomm® Dragonwing™ Q-8750 8.717 ms 1 - 351 MB NPU
Swin-Base QNN_DLC float Qualcomm® SA8295P 26.152 ms 1 - 344 MB NPU
Swin-Base QNN_DLC float Snapdragon® 8 Elite Extreme Gen 6 Mobile 6.235 ms 1 - 361 MB NPU
Swin-Base QNN_DLC float Snapdragon® 8 Elite Gen 6 Mobile 8.03 ms 0 - 353 MB NPU
Swin-Base QNN_DLC w8a16 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 6.896 ms 0 - 398 MB NPU
Swin-Base QNN_DLC w8a16 Snapdragon® 8 Elite For Galaxy Mobile 8.969 ms 0 - 389 MB NPU
Swin-Base QNN_DLC w8a16 Snapdragon® X2 Elite 7.504 ms 0 - 0 MB NPU
Swin-Base QNN_DLC w8a16 Snapdragon® X Elite 19.557 ms 0 - 0 MB NPU
Swin-Base QNN_DLC w8a16 Snapdragon® 8 Gen 3 Mobile 12.963 ms 0 - 512 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-8275 17.075 ms 0 - 3 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 18.623 ms 0 - 4 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® SA8775P 18.904 ms 0 - 413 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® SA8650P 18.904 ms 0 - 413 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® SA8255P 18.904 ms 0 - 413 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-9075 19.158 ms 2 - 5 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-X7181 19.557 ms 0 - 0 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-6690 113.75 ms 0 - 639 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-7790 20.758 ms 0 - 591 MB NPU
Swin-Base QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-8750 8.969 ms 0 - 389 MB NPU
Swin-Base QNN_DLC w8a16 Snapdragon® 7 Gen 4 Mobile 20.758 ms 0 - 591 MB NPU
Swin-Base QNN_DLC w8a16 Snapdragon® 8 Elite Extreme Gen 6 Mobile 6.587 ms 0 - 395 MB NPU
Swin-Base QNN_DLC w8a16 Snapdragon® 8 Elite Gen 6 Mobile 9.746 ms 0 - 391 MB NPU
Swin-Base TFLITE float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 7.459 ms 0 - 381 MB NPU
Swin-Base TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 9.193 ms 0 - 364 MB NPU
Swin-Base TFLITE float Snapdragon® 8 Gen 3 Mobile 12.818 ms 0 - 504 MB NPU
Swin-Base TFLITE float Snapdragon® 8 Gen 1 Mobile 27.511 ms 0 - 498 MB NPU
Swin-Base TFLITE float Qualcomm® Dragonwing™ IQ-8275 20.065 ms 0 - 179 MB NPU
Swin-Base TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 18.146 ms 0 - 4 MB NPU
Swin-Base TFLITE float Qualcomm® SA8775P 114.874 ms 0 - 34 MB GPU
Swin-Base TFLITE float Qualcomm® SA8650P 114.874 ms 0 - 34 MB GPU
Swin-Base TFLITE float Qualcomm® SA8255P 114.874 ms 0 - 34 MB GPU
Swin-Base TFLITE float Qualcomm® QCS8450 27.511 ms 0 - 498 MB NPU
Swin-Base TFLITE float Qualcomm® Dragonwing™ IQ-9075 21.194 ms 0 - 179 MB NPU
Swin-Base TFLITE float Qualcomm® Dragonwing™ Q-8750 9.193 ms 0 - 364 MB NPU
Swin-Base TFLITE float Qualcomm® SA7255P 268.492 ms 1 - 34 MB GPU
Swin-Base TFLITE float Qualcomm® SA8295P 26.844 ms 0 - 352 MB NPU
Swin-Base TFLITE float Snapdragon® 8 Elite Extreme Gen 6 Mobile 6.739 ms 0 - 368 MB NPU
Swin-Base TFLITE float Snapdragon® 8 Elite Gen 6 Mobile 8.699 ms 0 - 360 MB NPU

License

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

References

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Paper for qualcomm/Swin-Base