v0.42.0

See https://github.com/quic/ai-hub-models/releases/v0.42.0 for changelog.

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+DEPLOYMENT_MODEL_LICENSE.pdf filter=lfs diff=lfs merge=lfs -text
+SwinV2-Base_float.dlc filter=lfs diff=lfs merge=lfs -text

DEPLOYMENT_MODEL_LICENSE.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4409f93b0e82531303b3e10f52f1fdfb56467a25f05b7441c6bbd8bb8a64b42c
+size 109629

LICENSE

@@ -0,0 +1,2 @@
+The license of the original trained model can be found at https://github.com/pytorch/vision/blob/main/LICENSE.
+The license for the deployable model files (.tflite, .onnx, .dlc, .bin, etc.) can be found in DEPLOYMENT_MODEL_LICENSE.pdf.

README.md

@@ -0,0 +1,254 @@
+---
+library_name: pytorch
+license: other
+tags:
+- backbone
+- android
+pipeline_tag: image-classification
+
+---
+
+![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/swinv2_base/web-assets/model_demo.png)
+
+# SwinV2-Base: Optimized for Mobile Deployment
+## Imagenet classifier and general purpose backbone
+
+
+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 model is an implementation of SwinV2-Base found [here](https://github.com/pytorch/vision/blob/main/torchvision/models/swin_transformer.py).
+
+
+This repository provides scripts to run SwinV2-Base on Qualcomm® devices.
+More details on model performance across various devices, can be found
+[here](https://aihub.qualcomm.com/models/swinv2_base).
+
+
+
+### Model Details
+
+- **Model Type:** Model_use_case.image_classification
+- **Model Stats:**
+  - Model checkpoint: Imagenet
+  - Input resolution: 256x256
+  - Number of parameters: 88.8M
+  - Model size (float): 339 MB
+  - Model size (w8a16): 90.2 MB
+
+| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
+|---|---|---|---|---|---|---|---|---|
+| SwinV2-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 85.729 ms | 0 - 399 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 69.336 ms | 1 - 378 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 43.52 ms | 0 - 503 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 39.262 ms | 1 - 368 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 37.263 ms | 0 - 36 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 28.244 ms | 0 - 53 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 41.214 ms | 0 - 399 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 144.249 ms | 1 - 407 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 85.729 ms | 0 - 399 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 69.336 ms | 1 - 378 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 37.27 ms | 0 - 31 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 28.125 ms | 0 - 37 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 49.6 ms | 0 - 376 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 38.959 ms | 1 - 368 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 37.581 ms | 0 - 34 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 28.012 ms | 0 - 58 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 41.214 ms | 0 - 399 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 144.249 ms | 1 - 407 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 25.758 ms | 0 - 396 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 19.517 ms | 0 - 380 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 21.025 ms | 0 - 391 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 15.165 ms | 0 - 378 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 18.056 ms | 0 - 390 MB | NPU | [SwinV2-Base.tflite](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.tflite) |
+| SwinV2-Base | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 11.895 ms | 1 - 389 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+| SwinV2-Base | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 29.222 ms | 1033 - 1033 MB | NPU | [SwinV2-Base.dlc](https://huggingface.co/qualcomm/SwinV2-Base/blob/main/SwinV2-Base.dlc) |
+
+
+
+
+## Installation
+
+
+Install the package via pip:
+```bash
+pip install qai-hub-models
+```
+
+
+## Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device
+
+Sign-in to [Qualcomm® AI Hub Workbench](https://workbench.aihub.qualcomm.com/) with your
+Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
+
+With this API token, you can configure your client to run models on the cloud
+hosted devices.
+```bash
+qai-hub configure --api_token API_TOKEN
+```
+Navigate to [docs](https://workbench.aihub.qualcomm.com/docs/) for more information.
+
+
+
+## Demo off target
+
+The package contains a simple end-to-end demo that downloads pre-trained
+weights and runs this model on a sample input.
+
+```bash
+python -m qai_hub_models.models.swinv2_base.demo
+```
+
+The above demo runs a reference implementation of pre-processing, model
+inference, and post processing.
+
+**NOTE**: If you want running in a Jupyter Notebook or Google Colab like
+environment, please add the following to your cell (instead of the above).
+```
+%run -m qai_hub_models.models.swinv2_base.demo
+```
+
+
+### Run model on a cloud-hosted device
+
+In addition to the demo, you can also run the model on a cloud-hosted Qualcomm®
+device. This script does the following:
+* Performance check on-device on a cloud-hosted device
+* Downloads compiled assets that can be deployed on-device for Android.
+* Accuracy check between PyTorch and on-device outputs.
+
+```bash
+python -m qai_hub_models.models.swinv2_base.export
+```
+
+
+
+## How does this work?
+
+This [export script](https://aihub.qualcomm.com/models/swinv2_base/qai_hub_models/models/SwinV2-Base/export.py)
+leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) to optimize, validate, and deploy this model
+on-device. Lets go through each step below in detail:
+
+Step 1: **Compile model for on-device deployment**
+
+To compile a PyTorch model for on-device deployment, we first trace the model
+in memory using the `jit.trace` and then call the `submit_compile_job` API.
+
+```python
+import torch
+
+import qai_hub as hub
+from qai_hub_models.models.swinv2_base import Model
+
+# Load the model
+torch_model = Model.from_pretrained()
+
+# Device
+device = hub.Device("Samsung Galaxy S25")
+
+# Trace model
+input_shape = torch_model.get_input_spec()
+sample_inputs = torch_model.sample_inputs()
+
+pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
+
+# Compile model on a specific device
+compile_job = hub.submit_compile_job(
+    model=pt_model,
+    device=device,
+    input_specs=torch_model.get_input_spec(),
+)
+
+# Get target model to run on-device
+target_model = compile_job.get_target_model()
+
+```
+
+
+Step 2: **Performance profiling on cloud-hosted device**
+
+After compiling models from step 1. Models can be profiled model on-device using the
+`target_model`. Note that this scripts runs the model on a device automatically
+provisioned in the cloud.  Once the job is submitted, you can navigate to a
+provided job URL to view a variety of on-device performance metrics.
+```python
+profile_job = hub.submit_profile_job(
+    model=target_model,
+    device=device,
+)
+        
+```
+
+Step 3: **Verify on-device accuracy**
+
+To verify the accuracy of the model on-device, you can run on-device inference
+on sample input data on the same cloud hosted device.
+```python
+input_data = torch_model.sample_inputs()
+inference_job = hub.submit_inference_job(
+    model=target_model,
+    device=device,
+    inputs=input_data,
+)
+    on_device_output = inference_job.download_output_data()
+
+```
+With the output of the model, you can compute like PSNR, relative errors or
+spot check the output with expected output.
+
+**Note**: This on-device profiling and inference requires access to Qualcomm®
+AI Hub Workbench. [Sign up for access](https://myaccount.qualcomm.com/signup).
+
+
+
+## Run demo on a cloud-hosted device
+
+You can also run the demo on-device.
+
+```bash
+python -m qai_hub_models.models.swinv2_base.demo --eval-mode on-device
+```
+
+**NOTE**: If you want running in a Jupyter Notebook or Google Colab like
+environment, please add the following to your cell (instead of the above).
+```
+%run -m qai_hub_models.models.swinv2_base.demo -- --eval-mode on-device
+```
+
+
+## Deploying compiled model to Android
+
+
+The models can be deployed using multiple runtimes:
+- TensorFlow Lite (`.tflite` export): [This
+  tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a
+  guide to deploy the .tflite model in an Android application.
+
+
+- QNN (`.so` export ): This [sample
+  app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html)
+provides instructions on how to use the `.so` shared library  in an Android application.
+
+
+## View on Qualcomm® AI Hub
+Get more details on SwinV2-Base's performance across various devices [here](https://aihub.qualcomm.com/models/swinv2_base).
+Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
+
+
+## License
+* The license for the original implementation of SwinV2-Base can be found
+  [here](https://github.com/pytorch/vision/blob/main/LICENSE).
+* The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf)
+
+
+
+## References
+* [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883)
+* [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).
+
+

SwinV2-Base_float.dlc

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:91a3c1908cfd7bd91e6fb2e1be6d327fc755045f7fddf5560a7e09889dd3b582
+size 359463044

SwinV2-Base_float.tflite

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:413a52eff1da1c1a7b2815310ea8439c77154c4b754c2b31f953f635382dee70
+size 361828916

tool-versions.yaml

@@ -0,0 +1,3 @@
+tool_versions:
+  qnn_dlc:
+    qairt: 2.40.0.251030114326_189385