Hugging Face's logo

qualcomm
/
Facial-Attribute-Detection-Quantized

Object Detection
PyTorch
LiteRT
ONNX
real_time
quantized
android
Model card Files Community
Facial-Attribute-Detection-Quantized / README.md
qaihm-bot's picture
qaihm-bot
Upload README.md with huggingface_hub
0e87e6a verified
preview code
|
raw
history blame contribute delete
15.2 kB
metadata
library_name: pytorch
license: bsd-3-clause
tags:
  - real_time
  - quantized
  - android
pipeline_tag: object-detection

Facial-Attribute-Detection-Quantized: Optimized for Mobile Deployment

Comprehensive facial analysis by extracting face features

Facial feature extraction and additional attributes including liveness, eyeclose, mask and glasses detection for face recognition.

This model is an implementation of Facial-Attribute-Detection-Quantized found here.

This repository provides scripts to run Facial-Attribute-Detection-Quantized on Qualcomm® devices. More details on model performance across various devices, can be found here.

Model Details

  • Model Type: Object detection
  • Model Stats:
    • Model checkpoint: multitask_FR_state_dict.pt
    • Input resolution: 128x128
    • Input channel number: 1
    • Number of parameters: 11.6M
    • Model size: 47.6MB
Model Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Precision Primary Compute Unit Target Model
Facial-Attribute-Detection-Quantized Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 0.422 ms 0 - 50 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 0.418 ms 0 - 3 MB INT8 NPU Facial-Attribute-Detection-Quantized.so
Facial-Attribute-Detection-Quantized Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 0.64 ms 0 - 44 MB INT8 NPU Facial-Attribute-Detection-Quantized.onnx
Facial-Attribute-Detection-Quantized Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 0.323 ms 0 - 45 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 0.3 ms 0 - 21 MB INT8 NPU Facial-Attribute-Detection-Quantized.so
Facial-Attribute-Detection-Quantized Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 0.436 ms 0 - 54 MB INT8 NPU Facial-Attribute-Detection-Quantized.onnx
Facial-Attribute-Detection-Quantized Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 0.324 ms 0 - 35 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 0.327 ms 0 - 32 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 0.465 ms 0 - 43 MB INT8 NPU Facial-Attribute-Detection-Quantized.onnx
Facial-Attribute-Detection-Quantized SA7255P ADP SA7255P TFLITE 4.933 ms 0 - 31 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized SA7255P ADP SA7255P QNN 4.857 ms 0 - 10 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized SA8255 (Proxy) SA8255P Proxy TFLITE 0.427 ms 0 - 50 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized SA8255 (Proxy) SA8255P Proxy QNN 0.418 ms 0 - 3 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized SA8295P ADP SA8295P TFLITE 0.929 ms 0 - 34 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized SA8295P ADP SA8295P QNN 0.845 ms 0 - 18 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized SA8650 (Proxy) SA8650P Proxy TFLITE 0.419 ms 0 - 51 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized SA8650 (Proxy) SA8650P Proxy QNN 0.421 ms 0 - 3 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized SA8775P ADP SA8775P TFLITE 0.767 ms 0 - 31 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized SA8775P ADP SA8775P QNN 0.701 ms 0 - 10 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized RB3 Gen 2 (Proxy) QCS6490 Proxy TFLITE 1.46 ms 0 - 41 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized RB3 Gen 2 (Proxy) QCS6490 Proxy QNN 1.564 ms 0 - 14 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized RB5 (Proxy) QCS8250 Proxy TFLITE 69.358 ms 2 - 5 MB FP32 CPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized QCS8275 (Proxy) QCS8275 Proxy TFLITE 4.933 ms 0 - 31 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized QCS8275 (Proxy) QCS8275 Proxy QNN 4.857 ms 0 - 10 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized QCS8550 (Proxy) QCS8550 Proxy TFLITE 0.422 ms 0 - 49 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized QCS8550 (Proxy) QCS8550 Proxy QNN 0.417 ms 0 - 2 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized QCS9075 (Proxy) QCS9075 Proxy TFLITE 0.767 ms 0 - 31 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized QCS9075 (Proxy) QCS9075 Proxy QNN 0.701 ms 0 - 10 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized QCS8450 (Proxy) QCS8450 Proxy TFLITE 0.669 ms 0 - 45 MB INT8 NPU Facial-Attribute-Detection-Quantized.tflite
Facial-Attribute-Detection-Quantized QCS8450 (Proxy) QCS8450 Proxy QNN 0.65 ms 0 - 44 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized Snapdragon X Elite CRD Snapdragon® X Elite QNN 0.513 ms 0 - 0 MB INT8 NPU Use Export Script
Facial-Attribute-Detection-Quantized Snapdragon X Elite CRD Snapdragon® X Elite ONNX 0.609 ms 13 - 13 MB INT8 NPU Facial-Attribute-Detection-Quantized.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.face_attrib_net_quantized.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.face_attrib_net_quantized.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.face_attrib_net_quantized.export

Profiling Results
------------------------------------------------------------
Facial-Attribute-Detection-Quantized
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 0.4                    
Estimated peak memory usage (MB): [0, 50]                
Total # Ops                     : 171                    
Compute Unit(s)                 : NPU (171 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.face_attrib_net_quantized 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.face_attrib_net_quantized.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.face_attrib_net_quantized.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 Facial-Attribute-Detection-Quantized's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of Facial-Attribute-Detection-Quantized can be found here.
  • The license for the compiled assets for on-device deployment can be found here

References

Community