Beit: Optimized for Qualcomm Devices

Beit is a machine learning model that can classify images from the Imagenet dataset. It can also be used as a backbone in building more complex models for specific use cases.

This is based on the implementation of Beit 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 1.25.0	Download
ONNX	w8a16	Universal	QAIRT 2.42, ONNX Runtime 1.25.0	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a16	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

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

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

• Model checkpoint: Imagenet
• Input resolution: 224x224
• Number of parameters: 92.0M
• Model size (float): 351 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Beit	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	7.261 ms	1 - 466 MB	NPU
Beit	ONNX	float	Snapdragon® X2 Elite	7.42 ms	211 - 211 MB	NPU
Beit	ONNX	float	Snapdragon® X Elite	15.257 ms	183 - 183 MB	NPU
Beit	ONNX	float	Snapdragon® 8 Gen 3 Mobile	10.563 ms	1 - 509 MB	NPU
Beit	ONNX	float	Qualcomm® QCS8550 (Proxy)	14.683 ms	0 - 195 MB	NPU
Beit	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	8.629 ms	1 - 466 MB	NPU
Beit	ONNX	float	Qualcomm® QCS9075	18.148 ms	1 - 46 MB	NPU
Beit	ONNX	float	Qualcomm® QCS8750	8.629 ms	1 - 466 MB	NPU
Beit	ONNX	float	Qualcomm® QCS7181	15.257 ms	183 - 183 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	3.429 ms	0 - 275 MB	NPU
Beit	ONNX	w8a16	Snapdragon® X Elite	10.508 ms	149 - 149 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	6.963 ms	0 - 472 MB	NPU
Beit	ONNX	w8a16	Qualcomm® QCS6490	1070.497 ms	50 - 65 MB	CPU
Beit	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	10.19 ms	0 - 616 MB	NPU
Beit	ONNX	w8a16	Qualcomm® QCM6690	585.36 ms	66 - 77 MB	CPU
Beit	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	569.504 ms	67 - 79 MB	CPU
Beit	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	5.369 ms	0 - 385 MB	NPU
Beit	ONNX	w8a16	Qualcomm® QCS9075	10.715 ms	0 - 45 MB	NPU
Beit	ONNX	w8a16	Qualcomm® QCS7790	569.504 ms	67 - 79 MB	CPU
Beit	ONNX	w8a16	Qualcomm® QCS8750	5.369 ms	0 - 385 MB	NPU
Beit	ONNX	w8a16	Qualcomm® QCS7181	10.508 ms	149 - 149 MB	NPU
Beit	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	7.156 ms	1 - 299 MB	NPU
Beit	QNN_DLC	float	Snapdragon® X2 Elite	7.811 ms	1 - 1 MB	NPU
Beit	QNN_DLC	float	Snapdragon® X Elite	15.153 ms	1 - 1 MB	NPU
Beit	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	10.496 ms	0 - 394 MB	NPU
Beit	QNN_DLC	float	Qualcomm® QCS8275	47.996 ms	1 - 295 MB	NPU
Beit	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	14.431 ms	1 - 394 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA8775P	17.192 ms	1 - 295 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA8650P	17.192 ms	1 - 295 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA8255P	17.192 ms	1 - 295 MB	NPU
Beit	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	17.483 ms	0 - 381 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA7255P	47.996 ms	1 - 295 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA8295P	15.344 ms	1 - 289 MB	NPU
Beit	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	8.535 ms	1 - 299 MB	NPU
Beit	QNN_DLC	float	Qualcomm® QCS9075	18.361 ms	3 - 5 MB	NPU
Beit	QNN_DLC	float	Qualcomm® QCS8750	8.535 ms	1 - 299 MB	NPU
Beit	QNN_DLC	float	Qualcomm® QCS7181	15.153 ms	1 - 1 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	2.525 ms	0 - 244 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® X2 Elite	3.154 ms	0 - 0 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® X Elite	7.336 ms	0 - 0 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	4.66 ms	0 - 409 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® QCS8275	15.085 ms	0 - 352 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	6.774 ms	0 - 249 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® SA8775P	7.094 ms	0 - 352 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® SA8650P	7.094 ms	0 - 352 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® SA8255P	7.094 ms	0 - 352 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® QCM6690	71.605 ms	0 - 428 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® SA7255P	15.085 ms	0 - 352 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	8.609 ms	0 - 399 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	3.567 ms	0 - 349 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® QCS9075	7.779 ms	0 - 2 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® QCS7790	8.609 ms	0 - 399 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® QCS8750	3.567 ms	0 - 349 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® QCS7181	7.336 ms	0 - 0 MB	NPU
Beit	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	7.158 ms	0 - 306 MB	NPU
Beit	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	10.494 ms	0 - 400 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS8275	48.02 ms	0 - 300 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS8550 (Proxy)	14.093 ms	0 - 3 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8775P	17.196 ms	0 - 301 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8650P	17.196 ms	0 - 301 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8255P	17.196 ms	0 - 301 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS8450 (Proxy)	17.617 ms	0 - 369 MB	NPU
Beit	TFLITE	float	Qualcomm® SA7255P	48.02 ms	0 - 300 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8295P	15.324 ms	0 - 293 MB	NPU
Beit	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	8.508 ms	0 - 309 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS9075	18.356 ms	0 - 186 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS8750	8.508 ms	0 - 309 MB	NPU

License

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

References

• BEIT: BERT Pre-Training of Image Transformers
• 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.