Hugging Face's logo

qualcomm
/
CavaFace

Object Detection
PyTorch
LiteRT
real_time
android
Model card Files
xet
Community

CavaFace: Optimized for Mobile Deployment

Comprehensive facial analysis by extracting face features

A PyTorch-based framework for training face recognition models that generates facial embeddings for verification and identification tasks

This model is an implementation of CavaFace found here.

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

Model Details

  • Model Type: Model_use_case.object_detection
  • Model Stats:
    • Model checkpoint: IR_SE_100_Combined_Epoch_24.pt
    • Input resolution: 112x112
    • Number of parameters: 65.5M
    • Model size (float): 249.96MB
Model Precision Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit Target Model
CavaFace float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) TFLITE 24.767 ms 0 - 74 MB NPU CavaFace.tflite
CavaFace float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 24.739 ms 0 - 58 MB NPU CavaFace.dlc
CavaFace float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 7.121 ms 0 - 179 MB NPU CavaFace.tflite
CavaFace float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 8.871 ms 0 - 69 MB NPU CavaFace.dlc
CavaFace float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 4.401 ms 0 - 414 MB NPU CavaFace.tflite
CavaFace float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 4.311 ms 0 - 65 MB NPU CavaFace.dlc
CavaFace float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) ONNX 4.637 ms 0 - 12 MB NPU CavaFace.onnx.zip
CavaFace float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 7.086 ms 0 - 74 MB NPU CavaFace.tflite
CavaFace float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 30.228 ms 0 - 58 MB NPU CavaFace.dlc
CavaFace float SA7255P ADP Qualcomm® SA7255P TFLITE 24.767 ms 0 - 74 MB NPU CavaFace.tflite
CavaFace float SA7255P ADP Qualcomm® SA7255P QNN_DLC 24.739 ms 0 - 58 MB NPU CavaFace.dlc
CavaFace float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 4.34 ms 0 - 479 MB NPU CavaFace.tflite
CavaFace float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 4.315 ms 0 - 55 MB NPU CavaFace.dlc
CavaFace float SA8295P ADP Qualcomm® SA8295P TFLITE 8.217 ms 0 - 75 MB NPU CavaFace.tflite
CavaFace float SA8295P ADP Qualcomm® SA8295P QNN_DLC 7.973 ms 0 - 62 MB NPU CavaFace.dlc
CavaFace float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 4.349 ms 0 - 420 MB NPU CavaFace.tflite
CavaFace float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 4.318 ms 0 - 73 MB NPU CavaFace.dlc
CavaFace float SA8775P ADP Qualcomm® SA8775P TFLITE 7.086 ms 0 - 74 MB NPU CavaFace.tflite
CavaFace float SA8775P ADP Qualcomm® SA8775P QNN_DLC 30.228 ms 0 - 58 MB NPU CavaFace.dlc
CavaFace float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 3.228 ms 0 - 174 MB NPU CavaFace.tflite
CavaFace float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 3.206 ms 0 - 67 MB NPU CavaFace.dlc
CavaFace float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 3.377 ms 0 - 66 MB NPU CavaFace.onnx.zip
CavaFace float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile TFLITE 2.668 ms 0 - 79 MB NPU CavaFace.tflite
CavaFace float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile QNN_DLC 2.629 ms 0 - 65 MB NPU CavaFace.dlc
CavaFace float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile ONNX 2.792 ms 0 - 63 MB NPU CavaFace.onnx.zip
CavaFace float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile TFLITE 2.285 ms 0 - 80 MB NPU CavaFace.tflite
CavaFace float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile QNN_DLC 2.254 ms 0 - 64 MB NPU CavaFace.dlc
CavaFace float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile ONNX 2.446 ms 0 - 63 MB NPU CavaFace.onnx.zip
CavaFace float Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 4.473 ms 418 - 418 MB NPU CavaFace.dlc
CavaFace float Snapdragon X Elite CRD Snapdragon® X Elite ONNX 4.487 ms 127 - 127 MB NPU CavaFace.onnx.zip

Installation

Install the package via pip:

# NOTE: 3.10 <= PYTHON_VERSION < 3.14 is supported.
pip install "qai-hub-models[cavaface]"

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

Sign-in to Qualcomm® AI Hub Workbench 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.cavaface.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.cavaface.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.cavaface.export

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.cavaface 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.

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 Workbench. Sign up for access.

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 CavaFace's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

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

References

Community

Downloads last month
14
Inference Providers NEW
Object Detection
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support