Hugging Face's logo

qualcomm
/
VIT

Image Classification
PyTorch
LiteRT
backbone
android
Model card Files
xet
Community
VIT / README.md
qaihm-bot's picture
qaihm-bot
v0.32.0
cd90639 verified
preview code
|
raw
history blame
21.1 kB
 ---
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/vit/web-assets/model_demo.png)
# VIT: Optimized for Mobile Deployment
## Imagenet classifier and general purpose backbone
VIT 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 VIT found [here](https://github.com/pytorch/vision/blob/main/torchvision/models/vision_transformer.py).
This repository provides scripts to run VIT on Qualcomm® devices.
More details on model performance across various devices, can be found
[here](https://aihub.qualcomm.com/models/vit).
### Model Details
- **Model Type:** Model_use_case.image_classification
- **Model Stats:**
- Model checkpoint: Imagenet
- Input resolution: 224x224
- Number of parameters: 86.6M
- Model size (float): 330 MB
- Model size (w8a16): 86.2 MB
- Model size (w8a8): 83.2 MB
| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
|---|---|---|---|---|---|---|---|---|
| VIT | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 43.891 ms | 0 - 315 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 45.176 ms | 0 - 324 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 17.891 ms | 0 - 321 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 21.31 ms | 0 - 316 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 12.696 ms | 0 - 28 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 13.809 ms | 0 - 32 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 16.043 ms | 0 - 315 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 16.876 ms | 1 - 324 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 43.891 ms | 0 - 315 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 45.176 ms | 0 - 324 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 13.166 ms | 0 - 24 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 13.828 ms | 0 - 31 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 20.073 ms | 0 - 307 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 19.819 ms | 1 - 327 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 12.733 ms | 0 - 26 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 13.893 ms | 0 - 30 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 16.043 ms | 0 - 315 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 16.876 ms | 1 - 324 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 12.72 ms | 0 - 13 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 13.859 ms | 0 - 28 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 13.54 ms | 0 - 399 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT.onnx) |
| VIT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 8.994 ms | 0 - 319 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 9.587 ms | 38 - 370 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 9.308 ms | 10 - 344 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT.onnx) |
| VIT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 8.262 ms | 0 - 319 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 8.024 ms | 1 - 314 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 7.829 ms | 1 - 321 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT.onnx) |
| VIT | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 16.174 ms | 1116 - 1116 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 14.919 ms | 171 - 171 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT.onnx) |
| VIT | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 64.963 ms | 0 - 189 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 51.313 ms | 0 - 210 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 25.858 ms | 0 - 48 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 22.689 ms | 0 - 189 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 201.78 ms | 0 - 1634 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 64.963 ms | 0 - 189 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 25.98 ms | 0 - 48 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 37.12 ms | 0 - 212 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 26.02 ms | 0 - 47 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 22.689 ms | 0 - 189 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 26.113 ms | 0 - 48 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 150.726 ms | 578 - 827 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.onnx) |
| VIT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 19.328 ms | 0 - 195 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 125.044 ms | 673 - 835 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.onnx) |
| VIT | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 16.013 ms | 0 - 186 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 106.848 ms | 662 - 793 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.onnx) |
| VIT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 27.231 ms | 379 - 379 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 180.374 ms | 924 - 924 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.onnx) |
| VIT | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 24.381 ms | 0 - 48 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 30.273 ms | 0 - 164 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 12.796 ms | 0 - 57 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 16.349 ms | 0 - 227 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 12.099 ms | 3 - 117 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 10.505 ms | 0 - 27 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 12.516 ms | 0 - 49 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 9.592 ms | 0 - 164 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | TFLITE | 80.928 ms | 2 - 44 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 79.12 ms | 0 - 406 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 24.381 ms | 0 - 48 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 30.273 ms | 0 - 164 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 12.215 ms | 0 - 101 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 10.575 ms | 0 - 26 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 14.655 ms | 0 - 50 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 16.438 ms | 0 - 169 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 12.128 ms | 0 - 93 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 10.53 ms | 0 - 28 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 12.516 ms | 0 - 49 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 9.592 ms | 0 - 164 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 12.213 ms | 0 - 102 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 10.549 ms | 0 - 27 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 28.41 ms | 0 - 123 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.onnx) |
| VIT | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 8.69 ms | 0 - 53 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 6.985 ms | 0 - 167 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 20.049 ms | 0 - 278 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.onnx) |
| VIT | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 5.952 ms | 0 - 55 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 6.077 ms | 0 - 163 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 17.14 ms | 0 - 254 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.onnx) |
| VIT | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 11.728 ms | 425 - 425 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 34.308 ms | 88 - 88 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.onnx) |
## Installation
Install the package via pip:
```bash
pip install qai-hub-models
```
## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
Sign-in to [Qualcomm® AI Hub](https://app.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://app.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.vit.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.vit.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.vit.export
```
```
Profiling Results
------------------------------------------------------------
VIT
Device : cs_8275 (ANDROID 14)
Runtime : TFLITE
Estimated inference time (ms) : 43.9
Estimated peak memory usage (MB): [0, 315]
Total # Ops : 1579
Compute Unit(s) : npu (1579 ops) gpu (0 ops) cpu (0 ops)
```
## How does this work?
This [export script](https://aihub.qualcomm.com/models/vit/qai_hub_models/models/VIT/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.vit import Model
# Load the model
torch_model = Model.from_pretrained()
# Device
device = hub.Device("Samsung Galaxy S24")
# 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. [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.vit.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.vit.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 VIT's performance across various devices [here](https://aihub.qualcomm.com/models/vit).
Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
## License
* The license for the original implementation of VIT 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
* [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929)
* [Source Model Implementation](https://github.com/pytorch/vision/blob/main/torchvision/models/vision_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:[email protected]).