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 1.24.3	Download
ONNX	w8a8	Universal	QAIRT 2.42, ONNX Runtime 1.24.3	Download
QNN_DLC	float	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	12.147 ms	30 - 298 MB	NPU
PidNet	ONNX	float	Snapdragon® X2 Elite	13.032 ms	22 - 22 MB	NPU
PidNet	ONNX	float	Snapdragon® X Elite	32.73 ms	24 - 24 MB	NPU
PidNet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	23.118 ms	26 - 343 MB	NPU
PidNet	ONNX	float	Qualcomm® QCS8550 (Proxy)	32.643 ms	24 - 26 MB	NPU
PidNet	ONNX	float	Qualcomm® QCS9075	47.321 ms	24 - 51 MB	NPU
PidNet	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	15.98 ms	7 - 220 MB	NPU
PidNet	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	44.593 ms	73 - 293 MB	NPU
PidNet	ONNX	w8a8	Snapdragon® X2 Elite	45.557 ms	134 - 134 MB	NPU
PidNet	ONNX	w8a8	Snapdragon® X Elite	89.567 ms	133 - 133 MB	NPU
PidNet	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	45.633 ms	80 - 341 MB	NPU
PidNet	ONNX	w8a8	Qualcomm® QCS6490	386.898 ms	197 - 216 MB	CPU
PidNet	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	59.858 ms	101 - 109 MB	NPU
PidNet	ONNX	w8a8	Qualcomm® QCS9075	65.921 ms	103 - 106 MB	NPU
PidNet	ONNX	w8a8	Qualcomm® QCM6690	347.589 ms	141 - 151 MB	CPU
PidNet	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	44.093 ms	105 - 310 MB	NPU
PidNet	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	327.416 ms	134 - 145 MB	CPU
PidNet	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	11.992 ms	14 - 285 MB	NPU
PidNet	QNN_DLC	float	Snapdragon® X2 Elite	13.499 ms	24 - 24 MB	NPU
PidNet	QNN_DLC	float	Snapdragon® X Elite	38.895 ms	24 - 24 MB	NPU
PidNet	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	25.667 ms	24 - 326 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	117.69 ms	24 - 242 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	37.639 ms	25 - 27 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® SA8775P	46.618 ms	24 - 242 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS9075	60.965 ms	24 - 52 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	75.966 ms	6 - 311 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® SA7255P	117.69 ms	24 - 242 MB	NPU
PidNet	QNN_DLC	float	Qualcomm® SA8295P	51.942 ms	24 - 253 MB	NPU
PidNet	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	18.278 ms	23 - 269 MB	NPU
PidNet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	12.686 ms	2 - 275 MB	NPU
PidNet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	25.059 ms	0 - 319 MB	NPU
PidNet	TFLITE	float	Qualcomm® QCS8275 (Proxy)	116.587 ms	2 - 231 MB	NPU
PidNet	TFLITE	float	Qualcomm® QCS8550 (Proxy)	36.389 ms	2 - 6 MB	NPU
PidNet	TFLITE	float	Qualcomm® SA8775P	45.288 ms	2 - 233 MB	NPU
PidNet	TFLITE	float	Qualcomm® QCS9075	59.754 ms	0 - 45 MB	NPU
PidNet	TFLITE	float	Qualcomm® QCS8450 (Proxy)	74.773 ms	3 - 330 MB	NPU
PidNet	TFLITE	float	Qualcomm® SA7255P	116.587 ms	2 - 231 MB	NPU
PidNet	TFLITE	float	Qualcomm® SA8295P	50.984 ms	2 - 245 MB	NPU
PidNet	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	18.843 ms	2 - 256 MB	NPU
PidNet	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	22.222 ms	0 - 262 MB	NPU
PidNet	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	37.809 ms	1 - 267 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS6490	208.912 ms	3 - 73 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	98.277 ms	1 - 214 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	50.448 ms	1 - 4 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® SA8775P	51.011 ms	1 - 213 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS9075	52.894 ms	1 - 17 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCM6690	238.604 ms	1 - 233 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	58.791 ms	0 - 265 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® SA7255P	98.277 ms	1 - 214 MB	NPU
PidNet	TFLITE	w8a8	Qualcomm® SA8295P	58.058 ms	1 - 218 MB	NPU
PidNet	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	70.006 ms	1 - 237 MB	NPU
PidNet	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	66.96 ms	1 - 219 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