CavaFace: Optimized for Qualcomm Devices

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

This is based on the implementation of CavaFace 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 2.45, ONNX Runtime 1.25.0	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit CavaFace 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 CavaFace on GitHub for usage instructions.

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

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
CavaFace	ONNX	float	Snapdragon® X2 Elite	2.336 ms	212 - 212 MB	NPU
CavaFace	ONNX	float	Snapdragon® X Elite	4.312 ms	149 - 149 MB	NPU
CavaFace	ONNX	float	Snapdragon® 8 Gen 3 Mobile	3.207 ms	0 - 125 MB	NPU
CavaFace	ONNX	float	Snapdragon® 8 Gen 1 Mobile	9.514 ms	0 - 219 MB	NPU
CavaFace	ONNX	float	Qualcomm® QCS8550 (Proxy)	4.339 ms	0 - 132 MB	NPU
CavaFace	ONNX	float	Qualcomm® QCS8450	9.514 ms	0 - 219 MB	NPU
CavaFace	ONNX	float	Snapdragon® 8 Elite Mobile	2.628 ms	0 - 88 MB	NPU
CavaFace	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	2.252 ms	0 - 86 MB	NPU
CavaFace	ONNX	float	Qualcomm® QCS9075	6.751 ms	0 - 45 MB	NPU
CavaFace	ONNX	float	Qualcomm® QCS8750	2.628 ms	0 - 88 MB	NPU
CavaFace	ONNX	float	Qualcomm® QCS7181	4.312 ms	149 - 149 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® X2 Elite	2.551 ms	0 - 0 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® X Elite	4.493 ms	0 - 0 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	3.191 ms	0 - 129 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	9.589 ms	0 - 221 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS8275	24.707 ms	0 - 81 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	4.293 ms	0 - 269 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS8450	9.589 ms	0 - 221 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® 8 Elite Mobile	2.61 ms	0 - 83 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® SA8295P	8.006 ms	0 - 198 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	2.244 ms	0 - 83 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® SA7255P	24.707 ms	0 - 81 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS9075	6.733 ms	0 - 2 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS8750	2.61 ms	0 - 83 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS7181	4.493 ms	0 - 0 MB	NPU
CavaFace	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	3.144 ms	0 - 245 MB	NPU
CavaFace	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	9.461 ms	0 - 333 MB	NPU
CavaFace	TFLITE	float	Qualcomm® QCS8275	24.543 ms	0 - 95 MB	NPU
CavaFace	TFLITE	float	Qualcomm® QCS8550 (Proxy)	4.179 ms	0 - 2 MB	NPU
CavaFace	TFLITE	float	Qualcomm® SA8775P	68.557 ms	0 - 23 MB	GPU
CavaFace	TFLITE	float	Qualcomm® SA8650P	68.557 ms	0 - 23 MB	GPU
CavaFace	TFLITE	float	Qualcomm® SA8255P	68.557 ms	0 - 23 MB	GPU
CavaFace	TFLITE	float	Qualcomm® QCS8450	9.461 ms	0 - 333 MB	NPU
CavaFace	TFLITE	float	Snapdragon® 8 Elite Mobile	2.583 ms	0 - 96 MB	NPU
CavaFace	TFLITE	float	Qualcomm® SA8295P	7.95 ms	0 - 210 MB	NPU
CavaFace	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	2.224 ms	0 - 98 MB	NPU
CavaFace	TFLITE	float	Qualcomm® SA7255P	24.543 ms	0 - 95 MB	NPU
CavaFace	TFLITE	float	Qualcomm® QCS9075	6.638 ms	0 - 128 MB	NPU
CavaFace	TFLITE	float	Qualcomm® QCS8750	2.583 ms	0 - 96 MB	NPU

License

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

References

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.

Downloads last month: -; Downloads are not tracked for this model. How to track