Hugging Face's logo

qualcomm
/
Video-MAE

Video Classification
PyTorch
LiteRT
backbone
android
Model card Files
xet
Community
Video-MAE / README.md
qaihm-bot's picture
qaihm-bot
v0.45.0
f397dbf verified
preview code
|
raw
history blame contribute delete
11.4 kB
 ---
library_name: pytorch
license: other
tags:
- backbone
- android
pipeline_tag: video-classification
---
![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/video_mae/web-assets/model_demo.png)
# Video-MAE: Optimized for Mobile Deployment
## Sports and human action recognition in videos
Video MAE (Masked Auto Encoder) is a network for doing video classification that uses the ViT (Vision Transformer) backbone.
This model is an implementation of Video-MAE found [here](https://github.com/MCG-NJU/VideoMAE).
This repository provides scripts to run Video-MAE on Qualcomm® devices.
More details on model performance across various devices, can be found
[here](https://aihub.qualcomm.com/models/video_mae).
### Model Details
- **Model Type:** Model_use_case.video_classification
- **Model Stats:**
- Model checkpoint: Kinectics-400
- Input resolution: 224x224
- Number of parameters: 87.7M
- Model size (float): 335 MB
| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
|---|---|---|---|---|---|---|---|---|
| Video-MAE | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 423.568 ms | 0 - 908 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
| Video-MAE | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 1136.783 ms | 4 - 851 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 291.792 ms | 1 - 1054 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
| Video-MAE | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 575.304 ms | 9 - 1005 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 139.117 ms | 0 - 5 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
| Video-MAE | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 495.828 ms | 9 - 12 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 572.575 ms | 0 - 217 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
| Video-MAE | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 723.049 ms | 0 - 901 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
| Video-MAE | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 501.186 ms | 2 - 887 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 423.568 ms | 0 - 908 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
| Video-MAE | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 1136.783 ms | 4 - 851 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 669.5 ms | 0 - 845 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 723.049 ms | 0 - 901 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
| Video-MAE | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 501.186 ms | 2 - 887 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 100.899 ms | 1 - 1109 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
| Video-MAE | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 349.189 ms | 9 - 1059 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 416.581 ms | 1 - 1183 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
| Video-MAE | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 74.924 ms | 0 - 899 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
| Video-MAE | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 259.329 ms | 9 - 867 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 439.219 ms | 1 - 966 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
| Video-MAE | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | TFLITE | 59.82 ms | 1 - 913 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
| Video-MAE | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | QNN_DLC | 294.871 ms | 11 - 902 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | ONNX | 593.719 ms | 9 - 989 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
| Video-MAE | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 516.587 ms | 9 - 9 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
| Video-MAE | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 599.153 ms | 187 - 187 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
## Installation
Install the package via pip:
```bash
# NOTE: 3.10 <= PYTHON_VERSION < 3.14 is supported.
pip install "qai-hub-models[video-mae]"
```
## 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.video_mae.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.video_mae.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.video_mae.export
```
## How does this work?
This [export script](https://aihub.qualcomm.com/models/video_mae/qai_hub_models/models/Video-MAE/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.video_mae 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).
## 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 Video-MAE's performance across various devices [here](https://aihub.qualcomm.com/models/video_mae).
Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
## License
* The license for the original implementation of Video-MAE can be found
[here](https://github.com/MCG-NJU/VideoMAE/blob/main/LICENSE).
## References
* [Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602)
* [Source Model Implementation](https://github.com/MCG-NJU/VideoMAE)
## 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).