DDRNet23-Slim: Optimized for Qualcomm Devices

DDRNet23Slim is a machine learning model that segments an image into semantic classes, specifically designed for road-based scenes. It is designed for the application of self-driving cars.

This is based on the implementation of DDRNet23-Slim 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 1.27.1	Download
ONNX	w8a8	Universal	QAIRT 2.50, ONNX Runtime 1.27.1	Download
QNN_DLC	float	Universal	QAIRT 2.50	Download
QNN_DLC	w8a8	Universal	QAIRT 2.50	Download
TFLITE	float	Universal	QAIRT 2.50	Download

For more device-specific assets and performance metrics, visit DDRNet23-Slim 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 DDRNet23-Slim on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

Inference latency: RealTime
Input resolution: 2048x1024
Model checkpoint: DDRNet23s_imagenet.pth
Model size (float): 21.7 MB
Model size (w8a8): 6.11 MB
Number of output classes: 19
Number of parameters: 6.13M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
DDRNet23-Slim	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	11.11 ms	7 - 211 MB	NPU
DDRNet23-Slim	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	15.068 ms	6 - 184 MB	NPU
DDRNet23-Slim	ONNX	float	Snapdragon® X2 Elite	11.368 ms	24 - 24 MB	NPU
DDRNet23-Slim	ONNX	float	Snapdragon® X Elite	26.44 ms	24 - 24 MB	NPU
DDRNet23-Slim	ONNX	float	Snapdragon® 8 Gen 3 Mobile	19.767 ms	28 - 287 MB	NPU
DDRNet23-Slim	ONNX	float	Snapdragon® 8 Gen 1 Mobile	38.246 ms	1 - 264 MB	NPU
DDRNet23-Slim	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	33.903 ms	24 - 52 MB	NPU
DDRNet23-Slim	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	27.457 ms	24 - 40 MB	NPU
DDRNet23-Slim	ONNX	float	Qualcomm® QCS8450	38.246 ms	1 - 264 MB	NPU
DDRNet23-Slim	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	35.049 ms	24 - 51 MB	NPU
DDRNet23-Slim	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	26.44 ms	24 - 24 MB	NPU
DDRNet23-Slim	ONNX	float	Qualcomm® Dragonwing™ Q-8750	15.068 ms	6 - 184 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	26.616 ms	2 - 196 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	37.121 ms	1 - 190 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® X2 Elite	26.939 ms	7 - 7 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® X Elite	53.048 ms	7 - 7 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	37.46 ms	7 - 248 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® 8 Gen 1 Mobile	53.14 ms	8 - 244 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® Dragonwing™ QCS6490	109.927 ms	6 - 9 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-8275	49.087 ms	6 - 10 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	50.757 ms	6 - 16 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® QCS8450	53.14 ms	8 - 244 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-9075	52.462 ms	6 - 9 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-X7181	53.048 ms	7 - 7 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® Dragonwing™ Q-6690	197.647 ms	6 - 195 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® Dragonwing™ Q-7790	53.039 ms	6 - 202 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® Dragonwing™ Q-8750	37.121 ms	1 - 190 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	53.039 ms	6 - 202 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	10.987 ms	24 - 245 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	15.903 ms	15 - 221 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® X2 Elite	12.361 ms	24 - 24 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® X Elite	33.987 ms	24 - 24 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	22.551 ms	24 - 289 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	42.07 ms	24 - 290 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	38.889 ms	24 - 53 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	32.86 ms	24 - 26 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® QCS8450	42.07 ms	24 - 290 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	41.08 ms	24 - 53 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	33.987 ms	24 - 24 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	15.903 ms	15 - 221 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® SA8295P	43.4 ms	24 - 221 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	28.628 ms	6 - 228 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	39.505 ms	6 - 212 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® X2 Elite	29.18 ms	6 - 6 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® X Elite	59.12 ms	6 - 6 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	42.216 ms	6 - 245 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® 8 Gen 1 Mobile	56.923 ms	6 - 246 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-8275	54.908 ms	3 - 13 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	56.013 ms	6 - 118 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® QCS8450	56.923 ms	6 - 246 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-9075	57.794 ms	6 - 15 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-X7181	59.12 ms	6 - 6 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-8750	39.505 ms	6 - 212 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® SA8295P	64.112 ms	6 - 200 MB	NPU
DDRNet23-Slim	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	10.985 ms	2 - 227 MB	NPU
DDRNet23-Slim	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	15.862 ms	2 - 211 MB	NPU
DDRNet23-Slim	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	22.76 ms	2 - 272 MB	NPU
DDRNet23-Slim	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	42.355 ms	3 - 277 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	38.235 ms	0 - 42 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	33.108 ms	2 - 5 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® SA8775P	224.827 ms	3 - 32 MB	GPU
DDRNet23-Slim	TFLITE	float	Qualcomm® SA8650P	224.827 ms	3 - 32 MB	GPU
DDRNet23-Slim	TFLITE	float	Qualcomm® SA8255P	224.827 ms	3 - 32 MB	GPU
DDRNet23-Slim	TFLITE	float	Qualcomm® QCS8450	42.355 ms	3 - 277 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	41.314 ms	0 - 41 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	15.862 ms	2 - 211 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® SA7255P	324.979 ms	3 - 31 MB	GPU
DDRNet23-Slim	TFLITE	float	Qualcomm® SA8295P	43.193 ms	3 - 204 MB	NPU

License

The license for the original implementation of DDRNet23-Slim 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/DDRNet23-Slim

Deep Dual-resolution Networks for Real-time and Accurate Semantic Segmentation of Road Scenes

Paper • 2101.06085 • Published Jan 15, 2021