Sequencer2D: Optimized for Qualcomm Devices

Sequencer2D is a vision transformer model that can classify images from the Imagenet dataset.

This is based on the implementation of Sequencer2D 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 2.50, ONNX Runtime 1.30.0	Download
QNN_DLC	float	Universal	QAIRT 2.50	Download
TFLITE	float	Universal	QAIRT 2.50	Download
TFLITE	w8a8	Universal	QAIRT 2.50	Download

For more device-specific assets and performance metrics, visit Sequencer2D 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 Sequencer2D on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

• Input resolution: 224x224
• Model checkpoint: sequencer2d_s
• Model size (float): 106 MB
• Model size (w8a8): 69.1 MB
• Number of parameters: 27.6M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Sequencer2D	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	7.505 ms	0 - 757 MB	NPU
Sequencer2D	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	9.343 ms	0 - 729 MB	NPU
Sequencer2D	ONNX	float	Snapdragon® X2 Elite	8.663 ms	2 - 2 MB	NPU
Sequencer2D	ONNX	float	Snapdragon® X Elite	15.981 ms	60 - 60 MB	NPU
Sequencer2D	ONNX	float	Snapdragon® 8 Gen 3 Mobile	11.14 ms	0 - 844 MB	NPU
Sequencer2D	ONNX	float	Snapdragon® 8 Gen 1 Mobile	23.442 ms	1 - 736 MB	NPU
Sequencer2D	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	15.926 ms	1 - 5 MB	NPU
Sequencer2D	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	15.83 ms	0 - 67 MB	NPU
Sequencer2D	ONNX	float	Qualcomm® QCS8450	23.442 ms	1 - 736 MB	NPU
Sequencer2D	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	17.925 ms	1 - 4 MB	NPU
Sequencer2D	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	15.981 ms	60 - 60 MB	NPU
Sequencer2D	ONNX	float	Qualcomm® Dragonwing™ Q-8750	9.343 ms	0 - 729 MB	NPU
Sequencer2D	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	9.829 ms	0 - 915 MB	NPU
Sequencer2D	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	12.461 ms	1 - 871 MB	NPU
Sequencer2D	QNN_DLC	float	Snapdragon® X2 Elite	11.052 ms	1 - 1 MB	NPU
Sequencer2D	QNN_DLC	float	Snapdragon® X Elite	19.816 ms	1 - 1 MB	NPU
Sequencer2D	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	14.564 ms	0 - 991 MB	NPU
Sequencer2D	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	26.905 ms	0 - 866 MB	NPU
Sequencer2D	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	19.906 ms	1 - 4 MB	NPU
Sequencer2D	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	20.014 ms	1 - 478 MB	NPU
Sequencer2D	QNN_DLC	float	Qualcomm® SA8650P	24.036 ms	1 - 901 MB	NPU
Sequencer2D	QNN_DLC	float	Qualcomm® SA8255P	24.036 ms	1 - 901 MB	NPU
Sequencer2D	QNN_DLC	float	Qualcomm® QCS8450	26.905 ms	0 - 866 MB	NPU
Sequencer2D	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	21.881 ms	1 - 3 MB	NPU
Sequencer2D	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	19.816 ms	1 - 1 MB	NPU
Sequencer2D	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	12.461 ms	1 - 871 MB	NPU
Sequencer2D	QNN_DLC	float	Qualcomm® SA8295P	27.349 ms	1 - 740 MB	NPU
Sequencer2D	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	8.52 ms	0 - 889 MB	NPU
Sequencer2D	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	11.347 ms	0 - 857 MB	NPU
Sequencer2D	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	14.405 ms	0 - 983 MB	NPU
Sequencer2D	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	23.578 ms	0 - 720 MB	NPU
Sequencer2D	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	19.314 ms	0 - 80 MB	NPU
Sequencer2D	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	20.239 ms	0 - 9 MB	NPU
Sequencer2D	TFLITE	float	Qualcomm® SA8650P	19.048 ms	0 - 796 MB	NPU
Sequencer2D	TFLITE	float	Qualcomm® SA8255P	19.048 ms	0 - 796 MB	NPU
Sequencer2D	TFLITE	float	Qualcomm® QCS8450	23.578 ms	0 - 720 MB	NPU
Sequencer2D	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	21.051 ms	0 - 79 MB	NPU
Sequencer2D	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	11.347 ms	0 - 857 MB	NPU
Sequencer2D	TFLITE	float	Qualcomm® SA8295P	25.122 ms	0 - 591 MB	NPU
Sequencer2D	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	9.037 ms	0 - 953 MB	NPU
Sequencer2D	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	12.558 ms	0 - 896 MB	NPU
Sequencer2D	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	16.046 ms	0 - 1050 MB	NPU
Sequencer2D	TFLITE	w8a8	Snapdragon® 8 Gen 1 Mobile	24.751 ms	0 - 864 MB	NPU
Sequencer2D	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-8275	20.406 ms	0 - 69 MB	NPU
Sequencer2D	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	22.191 ms	0 - 12 MB	NPU
Sequencer2D	TFLITE	w8a8	Qualcomm® SA8650P	20.405 ms	0 - 862 MB	NPU
Sequencer2D	TFLITE	w8a8	Qualcomm® SA8255P	20.405 ms	0 - 862 MB	NPU
Sequencer2D	TFLITE	w8a8	Qualcomm® QCS8450	24.751 ms	0 - 864 MB	NPU
Sequencer2D	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-9075	22.136 ms	0 - 68 MB	NPU
Sequencer2D	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-8750	12.558 ms	0 - 896 MB	NPU
Sequencer2D	TFLITE	w8a8	Qualcomm® SA8295P	26.264 ms	0 - 705 MB	NPU

License

• The license for the original implementation of Sequencer2D can be found here.

References

• Sequencer: Deep LSTM for Image Classification
• 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.