README.md

---
library_name: pytorch
license: other
tags:
- android
pipeline_tag: keypoint-detection

---

![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/movenet/web-assets/model_demo.png)

# Movenet: Optimized for Qualcomm Devices

Movenet performs pose estimation on human images.

This is based on the implementation of Movenet found [here](https://github.com/lee-man/movenet-pytorch).
This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/movenet) 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 | ONNX Runtime 1.23.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/movenet/releases/v0.46.0/movenet-onnx-float.zip)
| ONNX | w8a16_mixed_int16 | Universal | ONNX Runtime 1.23.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/movenet/releases/v0.46.0/movenet-onnx-w8a16_mixed_int16.zip)
| QNN_DLC | float | Universal | QAIRT 2.42 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/movenet/releases/v0.46.0/movenet-qnn_dlc-float.zip)
| TFLITE | float | Universal | TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/movenet/releases/v0.46.0/movenet-tflite-float.zip)

For more device-specific assets and performance metrics, visit **[Movenet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/movenet)**.


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/movenet) 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 [Movenet on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/movenet) for usage instructions.

## Model Details

**Model Type:** Model_use_case.pose_estimation

**Model Stats:**
- Model checkpoint: None
- Input resolution: 192x192
- Number of parameters: 2.33M
- Model size (float): 8.91 MB

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| Movenet | ONNX | w8a16_mixed_int16 | Snapdragon® X Elite | 12.485 ms | 15 - 15 MB | CPU
| Movenet | ONNX | w8a16_mixed_int16 | Snapdragon® 8 Gen 3 Mobile | 13.131 ms | 14 - 29 MB | CPU
| Movenet | ONNX | w8a16_mixed_int16 | Qualcomm® QCS6490 | 57.324 ms | 13 - 16 MB | CPU
| Movenet | ONNX | w8a16_mixed_int16 | Qualcomm® QCS8550 (Proxy) | 14.479 ms | 12 - 22 MB | CPU
| Movenet | ONNX | w8a16_mixed_int16 | Qualcomm® QCS9075 | 22.717 ms | 13 - 16 MB | CPU
| Movenet | ONNX | w8a16_mixed_int16 | Qualcomm® QCM6690 | 28.585 ms | 15 - 24 MB | CPU
| Movenet | ONNX | w8a16_mixed_int16 | Snapdragon® 8 Elite For Galaxy Mobile | 10.267 ms | 15 - 24 MB | CPU
| Movenet | ONNX | w8a16_mixed_int16 | Snapdragon® 7 Gen 4 Mobile | 22.312 ms | 19 - 28 MB | CPU
| Movenet | ONNX | w8a16_mixed_int16 | Snapdragon® 8 Elite Gen 5 Mobile | 9.227 ms | 13 - 26 MB | CPU

## License
* The license for the original implementation of Movenet can be found
  [here](http://www.apache.org/licenses/LICENSE-2.0).

## References
* [MoveNet: Ultra fast and accurate pose detection model](https://blog.tensorflow.org/2021/05/next-generation-pose-detection-with-movenet-and-tensorflowjs.html)
* [Source Model Implementation](https://github.com/lee-man/movenet-pytorch)

## 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).