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 2.45, ONNX Runtime 1.27.1	Download
ONNX	w8a16	Universal	QAIRT 2.45, ONNX Runtime 1.27.1	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:

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

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Beit	ONNX	float	Snapdragon® X2 Elite	7.44 ms	2 - 2 MB	NPU
Beit	ONNX	float	Snapdragon® X Elite	15.399 ms	184 - 184 MB	NPU
Beit	ONNX	float	Snapdragon® 8 Gen 3 Mobile	10.531 ms	0 - 449 MB	NPU
Beit	ONNX	float	Snapdragon® 8 Gen 1 Mobile	17.916 ms	1 - 422 MB	NPU
Beit	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	17.891 ms	0 - 4 MB	NPU
Beit	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	14.87 ms	0 - 194 MB	NPU
Beit	ONNX	float	Qualcomm® QCS8450	17.916 ms	1 - 422 MB	NPU
Beit	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	17.923 ms	0 - 4 MB	NPU
Beit	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	15.399 ms	184 - 184 MB	NPU
Beit	ONNX	float	Qualcomm® Dragonwing™ Q-8750	8.548 ms	0 - 305 MB	NPU
Beit	ONNX	float	Snapdragon® 8 Elite Mobile	8.548 ms	0 - 305 MB	NPU
Beit	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	7.168 ms	0 - 305 MB	NPU
Beit	ONNX	w8a16	Snapdragon® X2 Elite	2.667 ms	1 - 1 MB	NPU
Beit	ONNX	w8a16	Snapdragon® X Elite	6.89 ms	95 - 95 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	4.579 ms	0 - 421 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	11.217 ms	0 - 424 MB	NPU
Beit	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	36.124 ms	0 - 3 MB	NPU
Beit	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-8275	6.266 ms	0 - 4 MB	NPU
Beit	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	6.703 ms	0 - 3 MB	NPU
Beit	ONNX	w8a16	Qualcomm® QCS8450	11.217 ms	0 - 424 MB	NPU
Beit	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	6.637 ms	0 - 3 MB	NPU
Beit	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	6.89 ms	95 - 95 MB	NPU
Beit	ONNX	w8a16	Qualcomm® Dragonwing™ Q-6690	64.048 ms	1 - 441 MB	NPU
Beit	ONNX	w8a16	Qualcomm® Dragonwing™ Q-7790	9.022 ms	0 - 409 MB	NPU
Beit	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	3.595 ms	0 - 361 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 8 Elite Mobile	3.595 ms	0 - 361 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	2.398 ms	0 - 256 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	9.022 ms	0 - 409 MB	NPU
Beit	QNN_DLC	float	Snapdragon® X2 Elite	7.859 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	10.476 ms	0 - 392 MB	NPU
Beit	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	17.563 ms	0 - 382 MB	NPU
Beit	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	17.676 ms	1 - 4 MB	NPU
Beit	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	14.357 ms	1 - 2 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA8775P	17.181 ms	1 - 296 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA8650P	17.181 ms	1 - 296 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA8255P	17.181 ms	1 - 296 MB	NPU
Beit	QNN_DLC	float	Qualcomm® QCS8450	17.563 ms	0 - 382 MB	NPU
Beit	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	17.467 ms	1 - 3 MB	NPU
Beit	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	15.153 ms	1 - 1 MB	NPU
Beit	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	8.492 ms	1 - 299 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA7255P	47.953 ms	1 - 295 MB	NPU
Beit	QNN_DLC	float	Qualcomm® SA8295P	15.316 ms	1 - 289 MB	NPU
Beit	QNN_DLC	float	Snapdragon® 8 Elite Mobile	8.492 ms	1 - 299 MB	NPU
Beit	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	7.191 ms	1 - 302 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® X2 Elite	3.194 ms	0 - 0 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® X Elite	7.344 ms	0 - 0 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	4.673 ms	0 - 411 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	6.288 ms	0 - 3 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	6.795 ms	0 - 2 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® SA8775P	7.044 ms	0 - 351 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® SA8650P	7.044 ms	0 - 351 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® SA8255P	7.044 ms	0 - 351 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	6.871 ms	0 - 2 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	7.344 ms	0 - 0 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-6690	64.704 ms	0 - 429 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-7790	8.61 ms	0 - 399 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	3.559 ms	0 - 351 MB	NPU
Beit	QNN_DLC	w8a16	Qualcomm® SA7255P	15.087 ms	0 - 351 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	3.559 ms	0 - 351 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	2.558 ms	0 - 247 MB	NPU
Beit	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	8.61 ms	0 - 399 MB	NPU
Beit	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	10.526 ms	0 - 406 MB	NPU
Beit	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	17.437 ms	0 - 372 MB	NPU
Beit	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	17.925 ms	0 - 187 MB	NPU
Beit	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	14.034 ms	0 - 2 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8775P	17.176 ms	0 - 301 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8650P	17.176 ms	0 - 301 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8255P	17.176 ms	0 - 301 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS8450	17.437 ms	0 - 372 MB	NPU
Beit	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	17.393 ms	0 - 186 MB	NPU
Beit	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	8.512 ms	0 - 310 MB	NPU
Beit	TFLITE	float	Qualcomm® SA7255P	47.986 ms	0 - 301 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8295P	15.342 ms	0 - 293 MB	NPU
Beit	TFLITE	float	Snapdragon® 8 Elite Mobile	8.512 ms	0 - 310 MB	NPU
Beit	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	7.182 ms	0 - 313 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.