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 1.24.3	Download
ONNX	w8a16	Universal	QAIRT 2.42, ONNX Runtime 1.24.3	Download
TFLITE	float	Universal	QAIRT 2.43, TFLite 2.19.1	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	3.739 ms	1 - 476 MB	NPU
Beit	ONNX	float	Snapdragon® X2 Elite	4.056 ms	185 - 185 MB	NPU
Beit	ONNX	float	Snapdragon® X Elite	9.973 ms	185 - 185 MB	NPU
Beit	ONNX	float	Snapdragon® 8 Gen 3 Mobile	6.282 ms	0 - 408 MB	NPU
Beit	ONNX	float	Qualcomm® QCS8550 (Proxy)	9.242 ms	0 - 196 MB	NPU
Beit	ONNX	float	Qualcomm® QCS9075	12.767 ms	0 - 4 MB	NPU
Beit	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	4.622 ms	0 - 483 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	3.772 ms	0 - 291 MB	NPU
Beit	ONNX	w8a16	Snapdragon® X2 Elite	4.104 ms	102 - 102 MB	NPU
Beit	ONNX	w8a16	Snapdragon® X Elite	10.979 ms	102 - 102 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	6.965 ms	0 - 409 MB	NPU
Beit	ONNX	w8a16	Qualcomm® QCS6490	1072.127 ms	39 - 337 MB	CPU
Beit	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	10.3 ms	0 - 123 MB	NPU
Beit	ONNX	w8a16	Qualcomm® QCS9075	10.654 ms	0 - 3 MB	NPU
Beit	ONNX	w8a16	Qualcomm® QCM6690	609.294 ms	96 - 111 MB	CPU
Beit	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	5.734 ms	0 - 401 MB	NPU
Beit	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	583.008 ms	99 - 115 MB	CPU
Beit	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	3.373 ms	0 - 465 MB	NPU
Beit	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	5.709 ms	0 - 518 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS8275 (Proxy)	36.48 ms	0 - 471 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS8550 (Proxy)	7.672 ms	0 - 3 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8775P	11.083 ms	0 - 471 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS9075	11.696 ms	0 - 186 MB	NPU
Beit	TFLITE	float	Qualcomm® QCS8450 (Proxy)	17.806 ms	0 - 475 MB	NPU
Beit	TFLITE	float	Qualcomm® SA7255P	36.48 ms	0 - 471 MB	NPU
Beit	TFLITE	float	Qualcomm® SA8295P	14.51 ms	1 - 455 MB	NPU
Beit	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	4.14 ms	0 - 477 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.