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 2.42, 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 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:

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

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 Mobile	10.453 ms	30 - 255 MB	NPU
DDRNet23-Slim	ONNX	float	Snapdragon® X2 Elite	10.95 ms	22 - 22 MB	NPU
DDRNet23-Slim	ONNX	float	Snapdragon® X Elite	28.071 ms	24 - 24 MB	NPU
DDRNet23-Slim	ONNX	float	Snapdragon® 8 Gen 3 Mobile	20.021 ms	30 - 310 MB	NPU
DDRNet23-Slim	ONNX	float	Qualcomm® QCS8550 (Proxy)	29.065 ms	24 - 27 MB	NPU
DDRNet23-Slim	ONNX	float	Qualcomm® QCS9075	39.815 ms	24 - 51 MB	NPU
DDRNet23-Slim	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	13.643 ms	5 - 204 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	44.338 ms	86 - 289 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® X2 Elite	43.735 ms	133 - 133 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® X Elite	60.315 ms	131 - 131 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	43.562 ms	35 - 294 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® QCS6490	301.464 ms	199 - 217 MB	CPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	57.111 ms	81 - 92 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® QCS9075	63.374 ms	86 - 89 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Qualcomm® QCM6690	264.691 ms	171 - 179 MB	CPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	42.889 ms	82 - 282 MB	NPU
DDRNet23-Slim	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	249.794 ms	191 - 199 MB	CPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	10.366 ms	14 - 250 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® X2 Elite	11.814 ms	24 - 24 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® X Elite	33.548 ms	24 - 24 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	22.494 ms	24 - 309 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	32.648 ms	24 - 26 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® SA8775P	40.248 ms	24 - 223 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® QCS9075	53.739 ms	24 - 52 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	67.005 ms	5 - 290 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Qualcomm® SA8295P	43.039 ms	24 - 229 MB	NPU
DDRNet23-Slim	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	15.461 ms	24 - 247 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	47.168 ms	6 - 238 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® X2 Elite	48.559 ms	6 - 6 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® X Elite	58.748 ms	6 - 6 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	41.802 ms	6 - 255 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	107.82 ms	6 - 205 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	56.165 ms	6 - 16 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® SA8775P	56.965 ms	6 - 205 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® QCS9075	59.882 ms	6 - 14 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	61.81 ms	6 - 254 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® SA7255P	107.82 ms	6 - 205 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Qualcomm® SA8295P	64.394 ms	6 - 208 MB	NPU
DDRNet23-Slim	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	40.526 ms	6 - 221 MB	NPU
DDRNet23-Slim	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	10.323 ms	2 - 242 MB	NPU
DDRNet23-Slim	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	22.472 ms	1 - 295 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® QCS8275 (Proxy)	98.309 ms	3 - 206 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® QCS8550 (Proxy)	32.456 ms	2 - 5 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® SA8775P	40.185 ms	2 - 206 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® QCS9075	53.755 ms	0 - 41 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® QCS8450 (Proxy)	67.21 ms	2 - 297 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® SA7255P	98.309 ms	3 - 206 MB	NPU
DDRNet23-Slim	TFLITE	float	Qualcomm® SA8295P	42.982 ms	2 - 216 MB	NPU
DDRNet23-Slim	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	15.355 ms	2 - 228 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	44.629 ms	0 - 230 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	36.982 ms	1 - 252 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Qualcomm® QCS6490	178.591 ms	9 - 78 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	95.075 ms	1 - 198 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	48.952 ms	0 - 159 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Qualcomm® SA8775P	49.626 ms	1 - 198 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Qualcomm® QCS9075	52.022 ms	0 - 15 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Qualcomm® QCM6690	216.815 ms	9 - 194 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	56.476 ms	0 - 249 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Qualcomm® SA7255P	95.075 ms	1 - 198 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Qualcomm® SA8295P	56.026 ms	1 - 202 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	65.286 ms	1 - 217 MB	NPU
DDRNet23-Slim	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	62.795 ms	10 - 209 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