PidNet: Optimized for Qualcomm Devices

PIDNet (Proportional-Integral-Derivative Network) is a real-time semantic segmentation model based on PID controllers

This is based on the implementation of PidNet 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 1.25.0	Download
ONNX	w8a8	Universal	QAIRT 2.42, ONNX Runtime 1.25.0	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a8	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download
TFLITE	w8a8	Universal	QAIRT 2.45	Download

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

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

Model checkpoint: PIDNet_S_Cityscapes_val.pt
Inference latency: RealTime
Input resolution: 1024x2048
Number of output classes: 19
Number of parameters: 8.06M
Model size (float): 29.1 MB
Model size (w8a8): 8.02 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
PidNet	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	12.127 ms	29 - 295 MB	NPU
PidNet	ONNX	float	Snapdragon® X2 Elite	13.164 ms	188 - 188 MB	NPU
PidNet	ONNX	float	Snapdragon® X Elite	32.422 ms	157 - 157 MB	NPU
PidNet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	23.163 ms	30 - 346 MB	NPU
PidNet	ONNX	float	Qualcomm® QCS8550 (Proxy)	33.105 ms	24 - 100 MB	NPU
PidNet	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	15.998 ms	7 - 225 MB	NPU
PidNet	ONNX	float	Qualcomm® QCS9075	47.482 ms	24 - 93 MB	NPU
PidNet	ONNX	float	Qualcomm® QCS8750	15.998 ms	7 - 225 MB	NPU
PidNet	ONNX	float	Qualcomm® QCS7181	32.422 ms	157 - 157 MB	NPU
PidNet	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	43.877 ms	7 - 238 MB	NPU
PidNet	ONNX	w8a8	Snapdragon® X2 Elite	44.167 ms	206 - 206 MB	NPU
PidNet	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	39.239 ms	7 - 274 MB	NPU
PidNet	ONNX	w8a8	Qualcomm® QCS6490	394.068 ms	196 - 215 MB	CPU
PidNet	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	52.407 ms	5 - 45 MB	NPU
PidNet	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	327.73 ms	185 - 195 MB	CPU
PidNet	ONNX	w8a8	Qualcomm® QCM6690	347.7 ms	199 - 209 MB	CPU
PidNet	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	41.767 ms	2 - 218 MB	NPU
PidNet	ONNX	w8a8	Qualcomm® QCS9075	55.072 ms	4 - 51 MB	NPU
PidNet	ONNX	w8a8	Qualcomm® QCS7790	327.73 ms	185 - 195 MB	CPU
PidNet	ONNX	w8a8	Qualcomm® QCS8750	41.767 ms	2 - 218 MB	NPU
PidNet	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	11.978 ms	8 - 277 MB	NPU
PidNet	QNN_DLC	float	Snapdragon® X2 Elite	13.605 ms	24 - 24 MB	NPU
PidNet	QNN_DLC	float	Snapdragon® X Elite	38.93 ms	24 - 24 MB	NPU
PidNet	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	25.741 ms	24 - 329 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS8275	117.755 ms	24 - 242 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	37.587 ms	24 - 311 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® SA8775P	46.603 ms	24 - 243 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® SA8650P	46.603 ms	24 - 243 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® SA8255P	46.603 ms	24 - 243 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	75.992 ms	4 - 310 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® SA8295P	51.902 ms	24 - 252 MB	NPU
PidNet	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	18.264 ms	24 - 267 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® SA7255P	117.755 ms	24 - 242 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS9075	61.657 ms	24 - 52 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS8750	18.264 ms	24 - 267 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS7181	38.93 ms	24 - 24 MB	NPU
PidNet	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	24.568 ms	6 - 269 MB	NPU
PidNet	QNN_DLC	w8a8	Snapdragon® X2 Elite	25.225 ms	6 - 6 MB	NPU
PidNet	QNN_DLC	w8a8	Snapdragon® X Elite	60.535 ms	6 - 6 MB	NPU
PidNet	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	43.132 ms	6 - 267 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® QCS8275	111.425 ms	6 - 219 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	57.416 ms	6 - 8 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® SA8775P	58.394 ms	6 - 219 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® SA8650P	58.394 ms	6 - 219 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® SA8255P	58.394 ms	6 - 219 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® SA7255P	111.425 ms	6 - 219 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® SA8295P	66.426 ms	6 - 223 MB	NPU
PidNet	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	43.971 ms	6 - 241 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® QCS9075	61.808 ms	6 - 14 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	64.249 ms	6 - 268 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® QCS8750	43.971 ms	6 - 241 MB	NPU
PidNet	QNN_DLC	w8a8	Qualcomm® QCS7181	60.535 ms	6 - 6 MB	NPU
PidNet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	12.696 ms	2 - 274 MB	NPU
PidNet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	24.884 ms	2 - 323 MB	NPU
PidNet	TFLITE	float	Qualcomm® QCS8275	116.702 ms	3 - 232 MB	NPU
PidNet	TFLITE	float	Qualcomm® QCS8550 (Proxy)	36.548 ms	3 - 5 MB	NPU
PidNet	TFLITE	float	Qualcomm® SA8775P	45.27 ms	2 - 232 MB	NPU
PidNet	TFLITE	float	Qualcomm® SA8650P	45.27 ms	2 - 232 MB	NPU
PidNet	TFLITE	float	Qualcomm® SA8255P	45.27 ms	2 - 232 MB	NPU
PidNet	TFLITE	float	Qualcomm® QCS8450 (Proxy)	74.126 ms	2 - 329 MB	NPU
PidNet	TFLITE	float	Qualcomm® SA8295P	51.059 ms	3 - 245 MB	NPU
PidNet	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	18.754 ms	0 - 249 MB	NPU
PidNet	TFLITE	float	Qualcomm® SA7255P	116.702 ms	3 - 232 MB	NPU
PidNet	TFLITE	float	Qualcomm® QCS9075	59.455 ms	0 - 45 MB	NPU
PidNet	TFLITE	float	Qualcomm® QCS8750	18.754 ms	0 - 249 MB	NPU
PidNet	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	22.152 ms	1 - 266 MB	NPU
PidNet	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	37.738 ms	1 - 264 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS6490	201.939 ms	2 - 72 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS8275	98.436 ms	1 - 213 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	50.173 ms	1 - 3 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® SA8775P	51.005 ms	1 - 214 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® SA8650P	51.005 ms	1 - 214 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® SA8255P	51.005 ms	1 - 214 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® SA7255P	98.436 ms	1 - 213 MB	NPU
PidNet	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	66.541 ms	5 - 224 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® SA8295P	58.099 ms	1 - 218 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCM6690	236.512 ms	2 - 233 MB	NPU
PidNet	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	69.579 ms	1 - 238 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS9075	53.455 ms	1 - 17 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	58.313 ms	1 - 266 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS7790	66.541 ms	5 - 224 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS8750	69.579 ms	1 - 238 MB	NPU

License

The license for the original implementation of PidNet 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/PidNet

PIDNet: A Real-time Semantic Segmentation Network Inspired by PID Controllers

Paper • 2206.02066 • Published Jun 4, 2022