PSPNet: Optimized for Qualcomm Devices

PSPNet (Pyramid Scene Parsing Network) is a semantic segmentation model that captures global context information by applying pyramid pooling modules. It is designed to improve scene understanding by aggregating contextual features at multiple scales.

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	w8a8	Universal	QAIRT 2.42, ONNX Runtime 1.24.3	Download
QNN_DLC	float	Universal	QAIRT 2.43	Download
QNN_DLC	w8a8	Universal	QAIRT 2.43	Download
TFLITE	float	Universal	QAIRT 2.43, TFLite 2.19.1	Download
TFLITE	w8a8	Universal	QAIRT 2.43, TFLite 2.19.1	Download

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

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

Model checkpoint: pspnet101_ade20k.pth
Input resolution: 1x3x473x473
Number of parameters: 65.7M
Model size (float): 251 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
PSPNet	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	323.881 ms	237 - 1004 MB	NPU
PSPNet	ONNX	float	Snapdragon® X2 Elite	428.119 ms	266 - 266 MB	NPU
PSPNet	ONNX	float	Snapdragon® X Elite	697.352 ms	529 - 529 MB	NPU
PSPNet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	512.477 ms	0 - 916 MB	NPU
PSPNet	ONNX	float	Qualcomm® QCS8550 (Proxy)	603.61 ms	128 - 289 MB	NPU
PSPNet	ONNX	float	Qualcomm® QCS9075	1759.332 ms	8 - 13 MB	NPU
PSPNet	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	334.737 ms	247 - 1004 MB	NPU
PSPNet	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	17.309 ms	73 - 340 MB	NPU
PSPNet	ONNX	w8a8	Snapdragon® X2 Elite	15.303 ms	131 - 131 MB	NPU
PSPNet	ONNX	w8a8	Snapdragon® X Elite	28.655 ms	133 - 133 MB	NPU
PSPNet	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	21.853 ms	71 - 493 MB	NPU
PSPNet	ONNX	w8a8	Qualcomm® QCS6490	3462.983 ms	203 - 282 MB	CPU
PSPNet	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	29.504 ms	64 - 147 MB	NPU
PSPNet	ONNX	w8a8	Qualcomm® QCS9075	34.109 ms	70 - 74 MB	NPU
PSPNet	ONNX	w8a8	Qualcomm® QCM6690	3036.562 ms	72 - 83 MB	CPU
PSPNet	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	19.166 ms	68 - 335 MB	NPU
PSPNet	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	2638.982 ms	52 - 63 MB	CPU
PSPNet	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	270.754 ms	3 - 742 MB	NPU
PSPNet	QNN_DLC	float	Snapdragon® X2 Elite	229.112 ms	3 - 3 MB	NPU
PSPNet	QNN_DLC	float	Snapdragon® X Elite	545.415 ms	3 - 3 MB	NPU
PSPNet	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	623.937 ms	0 - 873 MB	NPU
PSPNet	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	1869.395 ms	2 - 722 MB	NPU
PSPNet	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	602.875 ms	0 - 29 MB	NPU
PSPNet	QNN_DLC	float	Qualcomm® QCS9075	1754.111 ms	3 - 135 MB	NPU
PSPNet	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	1693.991 ms	4 - 436 MB	NPU
PSPNet	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	302.683 ms	0 - 716 MB	NPU
PSPNet	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	16.229 ms	1 - 238 MB	NPU
PSPNet	QNN_DLC	w8a8	Snapdragon® X2 Elite	14.409 ms	1 - 1 MB	NPU
PSPNet	QNN_DLC	w8a8	Snapdragon® X Elite	26.029 ms	1 - 1 MB	NPU
PSPNet	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	20.539 ms	1 - 363 MB	NPU
PSPNet	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	63.746 ms	1 - 263 MB	NPU
PSPNet	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	26.807 ms	1 - 4 MB	NPU
PSPNet	QNN_DLC	w8a8	Qualcomm® QCS9075	32.074 ms	1 - 35 MB	NPU
PSPNet	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	36.785 ms	0 - 359 MB	NPU
PSPNet	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	18.029 ms	1 - 230 MB	NPU
PSPNet	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	92.414 ms	1 - 458 MB	NPU
PSPNet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	328.56 ms	128 - 965 MB	NPU
PSPNet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	531.762 ms	127 - 1181 MB	NPU
PSPNet	TFLITE	float	Qualcomm® QCS8275 (Proxy)	1684.617 ms	102 - 937 MB	NPU
PSPNet	TFLITE	float	Qualcomm® QCS8550 (Proxy)	616.78 ms	128 - 131 MB	NPU
PSPNet	TFLITE	float	Qualcomm® QCS9075	1759.888 ms	0 - 272 MB	NPU
PSPNet	TFLITE	float	Qualcomm® QCS8450 (Proxy)	1577.73 ms	0 - 573 MB	NPU
PSPNet	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	979.912 ms	128 - 884 MB	NPU
PSPNet	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	17.4 ms	32 - 287 MB	NPU
PSPNet	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	27.222 ms	32 - 423 MB	NPU
PSPNet	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	74.363 ms	32 - 305 MB	NPU
PSPNet	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	35.063 ms	32 - 35 MB	NPU
PSPNet	TFLITE	w8a8	Qualcomm® QCS9075	37.657 ms	32 - 131 MB	NPU
PSPNet	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	53.752 ms	32 - 421 MB	NPU
PSPNet	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	22.731 ms	32 - 283 MB	NPU
PSPNet	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	370.171 ms	358 - 879 MB	NPU

License

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

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