SINet: Optimized for Qualcomm Devices

SINet is a machine learning model that is designed to segment people from close-up portrait images in real time.

This is based on the implementation of SINet 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
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a16	Universal	QAIRT 2.45	Download
QNN_DLC	w8a8	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

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

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

Input resolution: 224x224
Model checkpoint: SINet.pth
Model size (float): 415 KB
Model size (w8a8): 241 KB
Number of output classes: 2 (foreground / background)
Number of parameters: 91.9K

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
SINet	ONNX	float	Snapdragon® X2 Elite	0.765 ms	2 - 2 MB	NPU
SINet	ONNX	float	Snapdragon® X Elite	1.771 ms	2 - 2 MB	NPU
SINet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	1.119 ms	0 - 48 MB	NPU
SINet	ONNX	float	Snapdragon® 8 Gen 1 Mobile	2.037 ms	1 - 56 MB	NPU
SINet	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	1.833 ms	1 - 5 MB	NPU
SINet	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.681 ms	1 - 3 MB	NPU
SINet	ONNX	float	Qualcomm® QCS8450	2.037 ms	1 - 56 MB	NPU
SINet	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	1.905 ms	1 - 4 MB	NPU
SINet	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	1.771 ms	2 - 2 MB	NPU
SINet	ONNX	float	Qualcomm® Dragonwing™ Q-8750	0.839 ms	0 - 34 MB	NPU
SINet	ONNX	float	Snapdragon® 8 Elite Mobile	0.839 ms	0 - 34 MB	NPU
SINet	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	0.709 ms	0 - 37 MB	NPU
SINet	QNN_DLC	float	Snapdragon® X2 Elite	1.414 ms	1 - 1 MB	NPU
SINet	QNN_DLC	float	Snapdragon® X Elite	3.827 ms	1 - 1 MB	NPU
SINet	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	2.108 ms	0 - 46 MB	NPU
SINet	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	3.749 ms	0 - 56 MB	NPU
SINet	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	3.062 ms	1 - 4 MB	NPU
SINet	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	6.64 ms	1 - 34 MB	NPU
SINet	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	3.428 ms	1 - 2 MB	NPU
SINet	QNN_DLC	float	Qualcomm® SA8775P	3.906 ms	1 - 36 MB	NPU
SINet	QNN_DLC	float	Qualcomm® SA8650P	3.906 ms	1 - 36 MB	NPU
SINet	QNN_DLC	float	Qualcomm® SA8255P	3.906 ms	1 - 36 MB	NPU
SINet	QNN_DLC	float	Qualcomm® QCS8450	3.749 ms	0 - 56 MB	NPU
SINet	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	4.982 ms	1 - 4 MB	NPU
SINet	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	3.827 ms	1 - 1 MB	NPU
SINet	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	1.819 ms	0 - 33 MB	NPU
SINet	QNN_DLC	float	Qualcomm® SA7255P	6.64 ms	1 - 34 MB	NPU
SINet	QNN_DLC	float	Qualcomm® SA8295P	4.647 ms	1 - 36 MB	NPU
SINet	QNN_DLC	float	Snapdragon® 8 Elite Mobile	1.819 ms	0 - 33 MB	NPU
SINet	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	1.13 ms	1 - 37 MB	NPU
SINet	QNN_DLC	w8a16	Snapdragon® X2 Elite	3.505 ms	0 - 0 MB	NPU
SINet	QNN_DLC	w8a16	Snapdragon® X Elite	7.17 ms	0 - 0 MB	NPU
SINet	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	4.077 ms	0 - 50 MB	NPU
SINet	QNN_DLC	w8a16	Snapdragon® 8 Gen 1 Mobile	7.172 ms	0 - 54 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	5.448 ms	0 - 3 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	10.797 ms	0 - 37 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	6.575 ms	0 - 2 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® SA8775P	7.055 ms	0 - 40 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® SA8650P	7.055 ms	0 - 40 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® SA8255P	7.055 ms	0 - 40 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® QCS8450	7.172 ms	0 - 54 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	8.251 ms	0 - 3 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	7.17 ms	0 - 0 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	3.439 ms	0 - 41 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® SA7255P	10.797 ms	0 - 37 MB	NPU
SINet	QNN_DLC	w8a16	Qualcomm® SA8295P	8.141 ms	0 - 37 MB	NPU
SINet	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	3.439 ms	0 - 41 MB	NPU
SINet	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	3.214 ms	0 - 40 MB	NPU
SINet	QNN_DLC	w8a8	Snapdragon® X2 Elite	2.345 ms	0 - 0 MB	NPU
SINet	QNN_DLC	w8a8	Snapdragon® X Elite	4.713 ms	0 - 0 MB	NPU
SINet	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	2.905 ms	0 - 47 MB	NPU
SINet	QNN_DLC	w8a8	Snapdragon® 8 Gen 1 Mobile	4.63 ms	0 - 51 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-8275	4.034 ms	0 - 3 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-8275	7.754 ms	0 - 35 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	4.487 ms	0 - 2 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® SA8775P	4.694 ms	0 - 39 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® SA8650P	4.694 ms	0 - 39 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® SA8255P	4.694 ms	0 - 39 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® QCS8450	4.63 ms	0 - 51 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-9075	4.827 ms	0 - 2 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-X7181	4.713 ms	0 - 0 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-8750	2.364 ms	0 - 40 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® SA7255P	7.754 ms	0 - 35 MB	NPU
SINet	QNN_DLC	w8a8	Qualcomm® SA8295P	5.21 ms	0 - 35 MB	NPU
SINet	QNN_DLC	w8a8	Snapdragon® 8 Elite Mobile	2.364 ms	0 - 40 MB	NPU
SINet	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	2.139 ms	0 - 38 MB	NPU
SINet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	2.111 ms	0 - 45 MB	NPU
SINet	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	3.773 ms	0 - 53 MB	NPU
SINet	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	3.089 ms	1 - 5 MB	NPU
SINet	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	6.673 ms	1 - 35 MB	NPU
SINet	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	3.41 ms	0 - 2 MB	NPU
SINet	TFLITE	float	Qualcomm® SA8775P	3.915 ms	0 - 38 MB	NPU
SINet	TFLITE	float	Qualcomm® SA8650P	3.915 ms	0 - 38 MB	NPU
SINet	TFLITE	float	Qualcomm® SA8255P	3.915 ms	0 - 38 MB	NPU
SINet	TFLITE	float	Qualcomm® QCS8450	3.773 ms	0 - 53 MB	NPU
SINet	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	3.714 ms	1 - 4 MB	NPU
SINet	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	1.799 ms	0 - 36 MB	NPU
SINet	TFLITE	float	Qualcomm® SA7255P	6.673 ms	1 - 35 MB	NPU
SINet	TFLITE	float	Qualcomm® SA8295P	4.667 ms	1 - 35 MB	NPU
SINet	TFLITE	float	Snapdragon® 8 Elite Mobile	1.799 ms	0 - 36 MB	NPU
SINet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	1.138 ms	0 - 35 MB	NPU

License

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

References

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.

Downloads last month: -; Downloads are not tracked for this model. How to track

Paper for qualcomm/SINet

SINet: Extreme Lightweight Portrait Segmentation Networks with Spatial Squeeze Modules and Information Blocking Decoder

Paper • 1911.09099 • Published Nov 20, 2019