Hugging Face's logo

qualcomm
/
DDRNet23-Slim

Image Segmentation
PyTorch
real_time
android
Model card Files
xet
Community

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 2.37, ONNX Runtime 1.23.0 Download
ONNX w8a8 Universal QAIRT 2.37, ONNX Runtime 1.23.0 Download
QNN_DLC float Universal QAIRT 2.42 Download
QNN_DLC w8a8 Universal QAIRT 2.42 Download
TFLITE float Universal QAIRT 2.42, TFLite 2.17.0 Download
TFLITE w8a8 Universal QAIRT 2.42, TFLite 2.17.0 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® X Elite 24.369 ms 24 - 24 MB NPU
DDRNet23-Slim ONNX float Snapdragon® 8 Gen 3 Mobile 16.642 ms 32 - 253 MB NPU
DDRNet23-Slim ONNX float Qualcomm® QCS8550 (Proxy) 25.11 ms 5 - 241 MB NPU
DDRNet23-Slim ONNX float Qualcomm® QCS9075 39.885 ms 24 - 50 MB NPU
DDRNet23-Slim ONNX float Snapdragon® 8 Elite For Galaxy Mobile 12.623 ms 7 - 151 MB NPU
DDRNet23-Slim ONNX float Snapdragon® 8 Elite Gen 5 Mobile 8.933 ms 12 - 202 MB NPU
DDRNet23-Slim ONNX w8a8 Snapdragon® X Elite 77.961 ms 131 - 131 MB NPU
DDRNet23-Slim ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 57.668 ms 92 - 286 MB NPU
DDRNet23-Slim ONNX w8a8 Qualcomm® QCS6490 300.097 ms 198 - 216 MB CPU
DDRNet23-Slim ONNX w8a8 Qualcomm® QCS8550 (Proxy) 75.484 ms 80 - 343 MB NPU
DDRNet23-Slim ONNX w8a8 Qualcomm® QCS9075 63.411 ms 87 - 90 MB NPU
DDRNet23-Slim ONNX w8a8 Qualcomm® QCM6690 269.607 ms 159 - 168 MB CPU
DDRNet23-Slim ONNX w8a8 Snapdragon® 8 Elite For Galaxy Mobile 40.839 ms 85 - 229 MB NPU
DDRNet23-Slim ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 249.681 ms 138 - 147 MB CPU
DDRNet23-Slim ONNX w8a8 Snapdragon® 8 Elite Gen 5 Mobile 40.573 ms 55 - 198 MB NPU
DDRNet23-Slim QNN_DLC float Snapdragon® X Elite 33.867 ms 24 - 24 MB NPU
DDRNet23-Slim QNN_DLC float Snapdragon® 8 Gen 3 Mobile 22.522 ms 24 - 307 MB NPU
DDRNet23-Slim QNN_DLC float Qualcomm® QCS8275 (Proxy) 98.015 ms 24 - 222 MB NPU
DDRNet23-Slim QNN_DLC float Qualcomm® QCS8550 (Proxy) 33.275 ms 24 - 26 MB NPU
DDRNet23-Slim QNN_DLC float Qualcomm® SA8775P 40.398 ms 24 - 223 MB NPU
DDRNet23-Slim QNN_DLC float Qualcomm® QCS9075 53.448 ms 24 - 52 MB NPU
DDRNet23-Slim QNN_DLC float Qualcomm® QCS8450 (Proxy) 66.976 ms 23 - 307 MB NPU
DDRNet23-Slim QNN_DLC float Qualcomm® SA7255P 98.015 ms 24 - 222 MB NPU
DDRNet23-Slim QNN_DLC float Qualcomm® SA8295P 43.476 ms 24 - 230 MB NPU
DDRNet23-Slim QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 15.491 ms 19 - 243 MB NPU
DDRNet23-Slim QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 10.457 ms 24 - 260 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Snapdragon® X Elite 58.885 ms 6 - 6 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 41.962 ms 6 - 254 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Qualcomm® QCS8275 (Proxy) 108.13 ms 6 - 205 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Qualcomm® QCS8550 (Proxy) 56.332 ms 6 - 8 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Qualcomm® SA8775P 57.084 ms 6 - 206 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Qualcomm® QCS9075 59.885 ms 6 - 14 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Qualcomm® QCS8450 (Proxy) 61.244 ms 6 - 254 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Qualcomm® SA7255P 108.13 ms 6 - 205 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Qualcomm® SA8295P 64.395 ms 6 - 208 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Snapdragon® 8 Elite For Galaxy Mobile 40.58 ms 6 - 221 MB NPU
DDRNet23-Slim QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 Mobile 46.904 ms 6 - 239 MB NPU
DDRNet23-Slim TFLITE float Snapdragon® 8 Gen 3 Mobile 22.461 ms 1 - 294 MB NPU
DDRNet23-Slim TFLITE float Qualcomm® QCS8275 (Proxy) 97.957 ms 3 - 207 MB NPU
DDRNet23-Slim TFLITE float Qualcomm® QCS8550 (Proxy) 33.017 ms 2 - 6 MB NPU
DDRNet23-Slim TFLITE float Qualcomm® SA8775P 40.491 ms 3 - 206 MB NPU
DDRNet23-Slim TFLITE float Qualcomm® QCS9075 53.746 ms 0 - 40 MB NPU
DDRNet23-Slim TFLITE float Qualcomm® QCS8450 (Proxy) 66.844 ms 3 - 298 MB NPU
DDRNet23-Slim TFLITE float Qualcomm® SA7255P 97.957 ms 3 - 207 MB NPU
DDRNet23-Slim TFLITE float Qualcomm® SA8295P 43.415 ms 2 - 216 MB NPU
DDRNet23-Slim TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 15.648 ms 2 - 231 MB NPU
DDRNet23-Slim TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 10.411 ms 2 - 245 MB NPU
DDRNet23-Slim TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 36.995 ms 1 - 252 MB NPU
DDRNet23-Slim TFLITE w8a8 Qualcomm® QCS6490 172.506 ms 10 - 78 MB NPU
DDRNet23-Slim TFLITE w8a8 Qualcomm® QCS8275 (Proxy) 95.269 ms 1 - 198 MB NPU
DDRNet23-Slim TFLITE w8a8 Qualcomm® QCS8550 (Proxy) 48.779 ms 1 - 3 MB NPU
DDRNet23-Slim TFLITE w8a8 Qualcomm® SA8775P 49.617 ms 1 - 200 MB NPU
DDRNet23-Slim TFLITE w8a8 Qualcomm® QCS9075 51.434 ms 0 - 15 MB NPU
DDRNet23-Slim TFLITE w8a8 Qualcomm® QCM6690 218.866 ms 9 - 233 MB NPU
DDRNet23-Slim TFLITE w8a8 Qualcomm® QCS8450 (Proxy) 56.596 ms 0 - 250 MB NPU
DDRNet23-Slim TFLITE w8a8 Qualcomm® SA7255P 95.269 ms 1 - 198 MB NPU
DDRNet23-Slim TFLITE w8a8 Qualcomm® SA8295P 56.032 ms 1 - 202 MB NPU
DDRNet23-Slim TFLITE w8a8 Snapdragon® 8 Elite For Galaxy Mobile 64.905 ms 1 - 214 MB NPU
DDRNet23-Slim TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 65.053 ms 9 - 210 MB NPU
DDRNet23-Slim TFLITE w8a8 Snapdragon® 8 Elite Gen 5 Mobile 44.851 ms 1 - 231 MB NPU

License

  • The license for the original implementation of DDRNet23-Slim can be found here.

References

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
Image Segmentation
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Paper for qualcomm/DDRNet23-Slim