MobileFaceNet: Optimized for Qualcomm Devices

MobileFaceNet is an efficient CNN that maps a 112x112 face image to a compact 128-dimensional embedding. Two embeddings are compared via cosine similarity to determine whether they belong to the same person, achieving 99.48% accuracy on the LFW benchmark. The model uses depthwise-separable convolutions and inverted residual blocks (MobileNetV2-style) to stay under 1M parameters, making it well-suited for real-time face verification on mobile and edge devices. Trained with ArcFace loss on MS-Celeb-1M.

This is based on the implementation of MobileFaceNet found here. This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.

Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up 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 QAIRT 2.51, ONNX Runtime 1.30.0 Download
ONNX w8a16 Universal QAIRT 2.51, ONNX Runtime 1.30.0 Download
QNN_DLC float Universal QAIRT 2.51 Download
QNN_DLC w8a16 Universal QAIRT 2.51 Download
TFLITE float Universal QAIRT 2.51 Download

For more device-specific assets and performance metrics, visit MobileFaceNet on Qualcomm® AI Hub.

Option 2: Export with Custom Configurations

Use the Qualcomm® AI Hub Models 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 MobileFaceNet on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.object_detection

Model Stats:

  • Embedding dimension: 128
  • Input resolution: 112x112
  • Model checkpoint: mobilefacenet.pt
  • Model size (float): 4MB
  • Number of parameters: 1M

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
MobileFaceNet ONNX float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.527 ms 0 - 39 MB NPU
MobileFaceNet ONNX float Snapdragon® 8 Elite For Galaxy Mobile 0.541 ms 0 - 40 MB NPU
MobileFaceNet ONNX float Snapdragon® X2 Elite 0.642 ms 1 - 1 MB NPU
MobileFaceNet ONNX float Snapdragon® X Elite 1.023 ms 0 - 0 MB NPU
MobileFaceNet ONNX float Snapdragon® 8 Gen 3 Mobile 0.704 ms 0 - 57 MB NPU
MobileFaceNet ONNX float Snapdragon® 8 Gen 1 Mobile 1.45 ms 0 - 59 MB NPU
MobileFaceNet ONNX float Qualcomm® Dragonwing™ IQ-8275 1.351 ms 0 - 4 MB NPU
MobileFaceNet ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.991 ms 0 - 4 MB NPU
MobileFaceNet ONNX float Qualcomm® QCS8450 1.45 ms 0 - 59 MB NPU
MobileFaceNet ONNX float Qualcomm® Dragonwing™ IQ-9075 1.259 ms 0 - 4 MB NPU
MobileFaceNet ONNX float Qualcomm® Dragonwing™ IQ-X7181 1.023 ms 0 - 0 MB NPU
MobileFaceNet ONNX float Qualcomm® Dragonwing™ Q-8750 0.541 ms 0 - 40 MB NPU
MobileFaceNet ONNX float Snapdragon® 8 Elite Extreme Gen 6 Mobile 0.451 ms 0 - 37 MB NPU
MobileFaceNet ONNX float Snapdragon® 8 Elite Gen 6 Mobile 0.45 ms 0 - 38 MB NPU
MobileFaceNet ONNX w8a16 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.37 ms 0 - 48 MB NPU
MobileFaceNet ONNX w8a16 Snapdragon® 8 Elite For Galaxy Mobile 0.446 ms 0 - 52 MB NPU
MobileFaceNet ONNX w8a16 Snapdragon® X2 Elite 0.45 ms 1 - 1 MB NPU
MobileFaceNet ONNX w8a16 Snapdragon® X Elite 0.801 ms 0 - 0 MB NPU
MobileFaceNet ONNX w8a16 Snapdragon® 8 Gen 3 Mobile 0.564 ms 0 - 60 MB NPU
MobileFaceNet ONNX w8a16 Snapdragon® 8 Gen 1 Mobile 1.034 ms 0 - 64 MB NPU
MobileFaceNet ONNX w8a16 Qualcomm® Dragonwing™ QCS6490 2.361 ms 0 - 3 MB NPU
MobileFaceNet ONNX w8a16 Qualcomm® Dragonwing™ IQ-8275 0.753 ms 0 - 4 MB NPU
MobileFaceNet ONNX w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.776 ms 0 - 2 MB NPU
MobileFaceNet ONNX w8a16 Qualcomm® QCS8450 1.034 ms 0 - 64 MB NPU
MobileFaceNet ONNX w8a16 Qualcomm® Dragonwing™ IQ-9075 0.795 ms 0 - 4 MB NPU
MobileFaceNet ONNX w8a16 Qualcomm® Dragonwing™ IQ-X7181 0.801 ms 0 - 0 MB NPU
MobileFaceNet ONNX w8a16 Qualcomm® Dragonwing™ Q-6690 5.358 ms 0 - 159 MB NPU
MobileFaceNet ONNX w8a16 Qualcomm® Dragonwing™ Q-7790 0.924 ms 0 - 48 MB NPU
MobileFaceNet ONNX w8a16 Qualcomm® Dragonwing™ Q-8750 0.446 ms 0 - 52 MB NPU
MobileFaceNet ONNX w8a16 Snapdragon® 7 Gen 4 Mobile 0.924 ms 0 - 48 MB NPU
MobileFaceNet ONNX w8a16 Snapdragon® 8 Elite Extreme Gen 6 Mobile 0.282 ms 0 - 48 MB NPU
MobileFaceNet ONNX w8a16 Snapdragon® 8 Elite Gen 6 Mobile 0.344 ms 0 - 49 MB NPU
MobileFaceNet QNN_DLC float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.528 ms 0 - 36 MB NPU
MobileFaceNet QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 0.538 ms 0 - 39 MB NPU
MobileFaceNet QNN_DLC float Snapdragon® X2 Elite 0.761 ms 0 - 0 MB NPU
MobileFaceNet QNN_DLC float Snapdragon® X Elite 1.129 ms 0 - 0 MB NPU
MobileFaceNet QNN_DLC float Snapdragon® 8 Gen 3 Mobile 0.779 ms 0 - 52 MB NPU
MobileFaceNet QNN_DLC float Snapdragon® 8 Gen 1 Mobile 1.443 ms 0 - 56 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® Dragonwing™ IQ-8275 1.33 ms 0 - 3 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.969 ms 0 - 40 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® SA8775P 1.482 ms 0 - 36 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® SA8650P 1.482 ms 0 - 36 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® SA8255P 1.482 ms 0 - 36 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® QCS8450 1.443 ms 0 - 56 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 1.251 ms 0 - 3 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 1.129 ms 0 - 0 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® Dragonwing™ Q-8750 0.538 ms 0 - 39 MB NPU
MobileFaceNet QNN_DLC float Qualcomm® SA8295P 1.673 ms 0 - 33 MB NPU
MobileFaceNet QNN_DLC float Snapdragon® 8 Elite Extreme Gen 6 Mobile 0.444 ms 0 - 35 MB NPU
MobileFaceNet QNN_DLC float Snapdragon® 8 Elite Gen 6 Mobile 0.436 ms 0 - 36 MB NPU
MobileFaceNet QNN_DLC w8a16 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.37 ms 0 - 42 MB NPU
MobileFaceNet QNN_DLC w8a16 Snapdragon® 8 Elite For Galaxy Mobile 0.448 ms 0 - 41 MB NPU
MobileFaceNet QNN_DLC w8a16 Snapdragon® X2 Elite 0.572 ms 0 - 0 MB NPU
MobileFaceNet QNN_DLC w8a16 Snapdragon® X Elite 0.903 ms 0 - 0 MB NPU
MobileFaceNet QNN_DLC w8a16 Snapdragon® 8 Gen 3 Mobile 0.679 ms 0 - 53 MB NPU
MobileFaceNet QNN_DLC w8a16 Snapdragon® 8 Gen 1 Mobile 1.026 ms 0 - 57 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS6490 2.249 ms 0 - 2 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-8275 0.735 ms 0 - 3 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.773 ms 0 - 2 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® SA8650P 1.201 ms 0 - 43 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® SA8255P 1.201 ms 0 - 43 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® QCS8450 1.026 ms 0 - 57 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-9075 0.797 ms 0 - 2 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-X7181 0.903 ms 0 - 0 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-6690 5.365 ms 0 - 152 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-7790 0.929 ms 0 - 41 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-8750 0.448 ms 0 - 41 MB NPU
MobileFaceNet QNN_DLC w8a16 Qualcomm® SA8295P 1.233 ms 0 - 40 MB NPU
MobileFaceNet QNN_DLC w8a16 Snapdragon® 7 Gen 4 Mobile 0.929 ms 0 - 41 MB NPU
MobileFaceNet QNN_DLC w8a16 Snapdragon® 8 Elite Extreme Gen 6 Mobile 0.282 ms 0 - 42 MB NPU
MobileFaceNet QNN_DLC w8a16 Snapdragon® 8 Elite Gen 6 Mobile 0.326 ms 0 - 42 MB NPU
MobileFaceNet TFLITE float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 0.524 ms 0 - 37 MB NPU
MobileFaceNet TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 0.561 ms 0 - 35 MB NPU
MobileFaceNet TFLITE float Snapdragon® 8 Gen 3 Mobile 0.697 ms 0 - 52 MB NPU
MobileFaceNet TFLITE float Snapdragon® 8 Gen 1 Mobile 1.393 ms 0 - 56 MB NPU
MobileFaceNet TFLITE float Qualcomm® Dragonwing™ IQ-8275 1.343 ms 0 - 6 MB NPU
MobileFaceNet TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.961 ms 0 - 2 MB NPU
MobileFaceNet TFLITE float Qualcomm® SA8775P 1.476 ms 0 - 38 MB NPU
MobileFaceNet TFLITE float Qualcomm® SA8650P 1.476 ms 0 - 38 MB NPU
MobileFaceNet TFLITE float Qualcomm® SA8255P 1.476 ms 0 - 38 MB NPU
MobileFaceNet TFLITE float Qualcomm® QCS8450 1.393 ms 0 - 56 MB NPU
MobileFaceNet TFLITE float Qualcomm® Dragonwing™ IQ-9075 1.231 ms 0 - 6 MB NPU
MobileFaceNet TFLITE float Qualcomm® Dragonwing™ Q-8750 0.561 ms 0 - 35 MB NPU
MobileFaceNet TFLITE float Qualcomm® SA8295P 1.663 ms 0 - 35 MB NPU
MobileFaceNet TFLITE float Snapdragon® 8 Elite Extreme Gen 6 Mobile 0.444 ms 0 - 38 MB NPU
MobileFaceNet TFLITE float Snapdragon® 8 Elite Gen 6 Mobile 0.454 ms 0 - 36 MB NPU

License

  • The license for the original implementation of MobileFaceNet can be found here.

References

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support