Movenet: Optimized for Mobile Deployment

Perform accurate human pose estimation

Movenet performs pose estimation on human images.

This model is an implementation of Movenet found here.

This repository provides scripts to run Movenet on Qualcomm® devices. More details on model performance across various devices, can be found here.

Model Details

• Model Type: Model_use_case.pose_estimation
• Model Stats:
  • Model checkpoint: None
  • Input resolution: 192x192
  • Number of parameters: 3.31M
  • Model size: 9.2 MB

Model	Precision	Device	Chipset	Target Runtime	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit	Target Model
Movenet	float	QCS8275 (Proxy)	Qualcomm® QCS8275 (Proxy)	TFLITE	6.258 ms	1 - 10 MB	CPU	Movenet.tflite
Movenet	float	QCS8450 (Proxy)	Qualcomm® QCS8450 (Proxy)	TFLITE	4.69 ms	1 - 24 MB	CPU	Movenet.tflite
Movenet	float	QCS8450 (Proxy)	Qualcomm® QCS8450 (Proxy)	QNN	24.014 ms	13 - 36 MB	CPU	Use Export Script
Movenet	float	QCS8550 (Proxy)	Qualcomm® QCS8550 (Proxy)	TFLITE	3.765 ms	0 - 2 MB	CPU	Movenet.tflite
Movenet	float	QCS9075 (Proxy)	Qualcomm® QCS9075 (Proxy)	TFLITE	5.496 ms	1 - 15 MB	CPU	Movenet.tflite
Movenet	float	SA7255P ADP	Qualcomm® SA7255P	TFLITE	6.258 ms	1 - 10 MB	CPU	Movenet.tflite
Movenet	float	SA8255 (Proxy)	Qualcomm® SA8255P (Proxy)	TFLITE	4.718 ms	1 - 4 MB	CPU	Movenet.tflite
Movenet	float	SA8295P ADP	Qualcomm® SA8295P	TFLITE	3.848 ms	1 - 18 MB	CPU	Movenet.tflite
Movenet	float	SA8650 (Proxy)	Qualcomm® SA8650P (Proxy)	TFLITE	4.038 ms	1 - 4 MB	CPU	Movenet.tflite
Movenet	float	SA8775P ADP	Qualcomm® SA8775P	TFLITE	5.496 ms	1 - 15 MB	CPU	Movenet.tflite
Movenet	float	Samsung Galaxy S23	Snapdragon® 8 Gen 2 Mobile	TFLITE	5.565 ms	1 - 4 MB	CPU	Movenet.tflite
Movenet	float	Samsung Galaxy S23	Snapdragon® 8 Gen 2 Mobile	QNN	18.677 ms	13 - 15 MB	CPU	Use Export Script
Movenet	float	Samsung Galaxy S23	Snapdragon® 8 Gen 2 Mobile	ONNX	10.683 ms	7 - 18 MB	CPU	Movenet.onnx
Movenet	float	Samsung Galaxy S24	Snapdragon® 8 Gen 3 Mobile	TFLITE	3.437 ms	0 - 19 MB	CPU	Movenet.tflite
Movenet	float	Samsung Galaxy S24	Snapdragon® 8 Gen 3 Mobile	QNN	7.328 ms	12 - 33 MB	CPU	Use Export Script
Movenet	float	Samsung Galaxy S24	Snapdragon® 8 Gen 3 Mobile	ONNX	14.255 ms	5 - 28 MB	CPU	Movenet.onnx
Movenet	float	Snapdragon 8 Elite QRD	Snapdragon® 8 Elite Mobile	TFLITE	2.974 ms	1 - 15 MB	CPU	Movenet.tflite
Movenet	float	Snapdragon 8 Elite QRD	Snapdragon® 8 Elite Mobile	QNN	7.05 ms	12 - 27 MB	CPU	Use Export Script
Movenet	float	Snapdragon 8 Elite QRD	Snapdragon® 8 Elite Mobile	ONNX	8.358 ms	8 - 23 MB	CPU	Movenet.onnx
Movenet	float	Snapdragon X Elite CRD	Snapdragon® X Elite	ONNX	8.449 ms	18 - 18 MB	CPU	Movenet.onnx

Installation

Install the package via pip:

pip install qai-hub-models

Configure Qualcomm® AI Hub to run this model on a cloud-hosted device

Sign-in to Qualcomm® AI Hub 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.

qai-hub configure --api_token API_TOKEN

Navigate to 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.

python -m qai_hub_models.models.movenet.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.movenet.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.

python -m qai_hub_models.models.movenet.export

Profiling Results
------------------------------------------------------------
Movenet
Device                          : cs_8275 (ANDROID 14)                 
Runtime                         : TFLITE                               
Estimated inference time (ms)   : 6.3                                  
Estimated peak memory usage (MB): [1, 10]                              
Total # Ops                     : 152                                  
Compute Unit(s)                 : npu (0 ops) gpu (0 ops) cpu (152 ops)

How does this work?

This export script leverages Qualcomm® AI Hub 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.

import torch

import qai_hub as hub
from qai_hub_models.models.movenet 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.

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.

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.

Run demo on a cloud-hosted device

You can also run the demo on-device.

python -m qai_hub_models.models.movenet.demo --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.movenet.demo -- --on-device

Deploying compiled model to Android

The models can be deployed using multiple runtimes:

• TensorFlow Lite (.tflite export): This tutorial provides a guide to deploy the .tflite model in an Android application.

• QNN (.so export ): This sample app provides instructions on how to use the .so shared library in an Android application.

View on Qualcomm® AI Hub

Get more details on Movenet's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

• The license for the original implementation of Movenet can be found here.
• The license for the compiled assets for on-device deployment can be found here

References

• MoveNet: Ultra fast and accurate pose detection model
• Source Model Implementation

Community

• Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
• For questions or feedback please reach out to us.