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 2.42, ONNX Runtime 1.24.1	Download
QNN_DLC	float	Universal	QAIRT 2.43	Download
TFLITE	float	Universal	QAIRT 2.43, TFLite 2.17.0	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	2.337 ms	126 - 126 MB	NPU
CavaFace	ONNX	float	Snapdragon® X Elite	4.511 ms	126 - 126 MB	NPU
CavaFace	ONNX	float	Snapdragon® 8 Gen 3 Mobile	3.2 ms	0 - 110 MB	NPU
CavaFace	ONNX	float	Qualcomm® QCS8550 (Proxy)	4.34 ms	0 - 131 MB	NPU
CavaFace	ONNX	float	Qualcomm® QCS9075	6.789 ms	0 - 3 MB	NPU
CavaFace	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	2.64 ms	0 - 79 MB	NPU
CavaFace	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	2.265 ms	0 - 92 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® X2 Elite	2.602 ms	0 - 0 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® X Elite	4.456 ms	0 - 0 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	3.187 ms	0 - 102 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	24.701 ms	0 - 81 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	4.299 ms	0 - 2 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® SA8775P	30.193 ms	0 - 81 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS9075	6.772 ms	0 - 2 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	8.941 ms	0 - 109 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® SA7255P	24.701 ms	0 - 81 MB	NPU
CavaFace	QNN_DLC	float	Qualcomm® SA8295P	7.956 ms	0 - 86 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	2.63 ms	0 - 85 MB	NPU
CavaFace	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	2.235 ms	0 - 84 MB	NPU
CavaFace	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	3.143 ms	0 - 213 MB	NPU
CavaFace	TFLITE	float	Qualcomm® QCS8275 (Proxy)	24.552 ms	0 - 98 MB	NPU
CavaFace	TFLITE	float	Qualcomm® QCS8550 (Proxy)	4.172 ms	0 - 2 MB	NPU
CavaFace	TFLITE	float	Qualcomm® SA8775P	6.896 ms	0 - 98 MB	NPU
CavaFace	TFLITE	float	Qualcomm® QCS9075	6.699 ms	0 - 129 MB	NPU
CavaFace	TFLITE	float	Qualcomm® QCS8450 (Proxy)	8.752 ms	0 - 221 MB	NPU
CavaFace	TFLITE	float	Qualcomm® SA7255P	24.552 ms	0 - 98 MB	NPU
CavaFace	TFLITE	float	Qualcomm® SA8295P	7.934 ms	0 - 101 MB	NPU
CavaFace	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	2.589 ms	0 - 102 MB	NPU
CavaFace	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	2.233 ms	0 - 101 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.