--- library_name: pytorch license: other tags: - backbone - bu_auto - android pipeline_tag: image-classification --- ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/swin_base/web-assets/model_demo.png) # 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](https://github.com/pytorch/vision/blob/main/torchvision/models/swin_transformer.py). This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/swin_base) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary). Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) 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](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/swin_base/releases/v0.54.0/swin_base-onnx-float.zip) | ONNX | w8a16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.3 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/swin_base/releases/v0.54.0/swin_base-onnx-w8a16.zip) | QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/swin_base/releases/v0.54.0/swin_base-qnn_dlc-float.zip) | QNN_DLC | w8a16 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/swin_base/releases/v0.54.0/swin_base-qnn_dlc-w8a16.zip) | TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/swin_base/releases/v0.54.0/swin_base-tflite-float.zip) For more device-specific assets and performance metrics, visit **[Swin-Base on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/swin_base)**. ### Option 2: Export with Custom Configurations Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/swin_base) 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](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/swin_base) for usage instructions. ## Model Details **Model Type:** Model_use_case.image_classification **Model Stats:** - Model checkpoint: Imagenet - Input resolution: 224x224 - Number of parameters: 88.8M - Model size (float): 339 MB - Model size (w8a16): 90.2 MB ## 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 Mobile | 7.842 ms | 1 - 534 MB | NPU | Swin-Base | ONNX | float | Snapdragon® 8 Elite Mobile | 9.783 ms | 0 - 511 MB | NPU | Swin-Base | ONNX | float | Snapdragon® X2 Elite | 8.376 ms | 175 - 175 MB | NPU | Swin-Base | ONNX | float | Snapdragon® X Elite | 19.682 ms | 174 - 174 MB | NPU | Swin-Base | ONNX | float | Snapdragon® X Elite | 19.682 ms | 174 - 174 MB | NPU | Swin-Base | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 13.052 ms | 1 - 608 MB | NPU | Swin-Base | ONNX | float | Qualcomm® QCS8550 (Proxy) | 18.905 ms | 0 - 196 MB | NPU | Swin-Base | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 9.783 ms | 0 - 511 MB | NPU | Swin-Base | ONNX | float | Qualcomm® QCS9075 | 24.452 ms | 0 - 4 MB | NPU | Swin-Base | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 6.969 ms | 0 - 448 MB | NPU | Swin-Base | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 8.785 ms | 0 - 424 MB | NPU | Swin-Base | ONNX | w8a16 | Snapdragon® X2 Elite | 7.518 ms | 92 - 92 MB | NPU | Swin-Base | ONNX | w8a16 | Snapdragon® X Elite | 17.605 ms | 91 - 91 MB | NPU | Swin-Base | ONNX | w8a16 | Snapdragon® X Elite | 17.605 ms | 91 - 91 MB | NPU | Swin-Base | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 11.385 ms | 0 - 554 MB | NPU | Swin-Base | ONNX | w8a16 | Qualcomm® QCS6490 | 1114.666 ms | 62 - 92 MB | CPU | Swin-Base | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 16.751 ms | 0 - 112 MB | NPU | Swin-Base | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 8.785 ms | 0 - 424 MB | NPU | Swin-Base | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 592.566 ms | 129 - 151 MB | CPU | Swin-Base | ONNX | w8a16 | Qualcomm® QCM6690 | 627.859 ms | 129 - 151 MB | CPU | Swin-Base | ONNX | w8a16 | Qualcomm® QCS9075 | 20.874 ms | 0 - 3 MB | NPU | Swin-Base | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 592.566 ms | 129 - 151 MB | CPU | Swin-Base | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 7.481 ms | 1 - 382 MB | NPU | Swin-Base | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 9.423 ms | 1 - 366 MB | NPU | Swin-Base | QNN_DLC | float | Snapdragon® X2 Elite | 8.482 ms | 1 - 1 MB | NPU | Swin-Base | QNN_DLC | float | Snapdragon® X Elite | 19.219 ms | 1 - 1 MB | NPU | Swin-Base | QNN_DLC | float | Snapdragon® X Elite | 19.219 ms | 1 - 1 MB | NPU | Swin-Base | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 12.409 ms | 0 - 515 MB | NPU | Swin-Base | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 18.341 ms | 1 - 2 MB | NPU | Swin-Base | QNN_DLC | float | Qualcomm® SA8775P | 21.382 ms | 1 - 362 MB | NPU | Swin-Base | QNN_DLC | float | Qualcomm® SA8775P | 21.382 ms | 1 - 362 MB | NPU | Swin-Base | QNN_DLC | float | Qualcomm® SA8775P | 21.382 ms | 1 - 362 MB | NPU | Swin-Base | QNN_DLC | float | Qualcomm® SA7255P | 53.592 ms | 1 - 362 MB | NPU | Swin-Base | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 29.1 ms | 0 - 505 MB | NPU | Swin-Base | QNN_DLC | float | Qualcomm® SA8295P | 27.631 ms | 1 - 353 MB | NPU | Swin-Base | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 9.423 ms | 1 - 366 MB | NPU | Swin-Base | QNN_DLC | float | Qualcomm® QCS9075 | 23.277 ms | 1 - 3 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 7.546 ms | 0 - 417 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 9.567 ms | 0 - 406 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 8.441 ms | 0 - 0 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Snapdragon® X Elite | 20.32 ms | 0 - 0 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Snapdragon® X Elite | 20.32 ms | 0 - 0 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 12.838 ms | 0 - 508 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 19.074 ms | 0 - 349 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Qualcomm® SA8775P | 19.545 ms | 0 - 411 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Qualcomm® SA8775P | 19.545 ms | 0 - 411 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Qualcomm® SA8775P | 19.545 ms | 0 - 411 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Qualcomm® SA7255P | 35.255 ms | 0 - 411 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 9.567 ms | 0 - 406 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 21.538 ms | 0 - 610 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 119.434 ms | 0 - 913 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 22.716 ms | 0 - 2 MB | NPU | Swin-Base | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 21.538 ms | 0 - 610 MB | NPU | Swin-Base | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 7.881 ms | 0 - 393 MB | NPU | Swin-Base | TFLITE | float | Snapdragon® 8 Elite Mobile | 9.681 ms | 0 - 379 MB | NPU | Swin-Base | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 12.779 ms | 0 - 1043 MB | NPU | Swin-Base | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 18.916 ms | 0 - 4 MB | NPU | Swin-Base | TFLITE | float | Qualcomm® SA8775P | 21.917 ms | 0 - 375 MB | NPU | Swin-Base | TFLITE | float | Qualcomm® SA8775P | 21.917 ms | 0 - 375 MB | NPU | Swin-Base | TFLITE | float | Qualcomm® SA8775P | 21.917 ms | 0 - 375 MB | NPU | Swin-Base | TFLITE | float | Qualcomm® SA7255P | 53.453 ms | 0 - 709 MB | NPU | Swin-Base | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 29.021 ms | 0 - 509 MB | NPU | Swin-Base | TFLITE | float | Qualcomm® SA8295P | 28.384 ms | 0 - 371 MB | NPU | Swin-Base | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 9.681 ms | 0 - 379 MB | NPU | Swin-Base | TFLITE | float | Qualcomm® QCS9075 | 23.89 ms | 0 - 178 MB | NPU ## License * The license for the original implementation of Swin-Base can be found [here](https://github.com/pytorch/vision/blob/main/LICENSE). ## References * [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) * [Source Model Implementation](https://github.com/pytorch/vision/blob/main/torchvision/models/swin_transformer.py) ## Community * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI. * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).