Image Segmentation
PyTorch
bu_auto
android

DeepLabXception: Optimized for Qualcomm Devices

DeepLabXception is a semantic segmentation model supporting multiple backbones like ResNet-101 and Xception, with flexible dataset compatibility including COCO, VOC, and Cityscapes.

This is based on the implementation of DeepLabXception 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 w8a16 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 w8a16 Universal QAIRT 2.50 Download
QNN_DLC w8a8 Universal QAIRT 2.50 Download
TFLITE float Universal QAIRT 2.50 Download
TFLITE w8a8 Universal QAIRT 2.50 Download

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

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

  • Input resolution: 480x520
  • Model checkpoint: COCO_WITH_VOC_LABELS_V1
  • Model size (float): 158 MB
  • Number of output classes: 21
  • Number of parameters: 41.26M

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
DeepLabXception ONNX float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 9.129 ms 1 - 248 MB NPU
DeepLabXception ONNX float Snapdragon® 8 Elite For Galaxy Mobile 11.892 ms 2 - 241 MB NPU
DeepLabXception ONNX float Snapdragon® X2 Elite 9.952 ms 3 - 3 MB NPU
DeepLabXception ONNX float Snapdragon® X Elite 21.325 ms 85 - 85 MB NPU
DeepLabXception ONNX float Snapdragon® 8 Gen 3 Mobile 15.428 ms 0 - 344 MB NPU
DeepLabXception ONNX float Snapdragon® 8 Gen 1 Mobile 36.87 ms 4 - 344 MB NPU
DeepLabXception ONNX float Qualcomm® Dragonwing™ IQ-8275 33.479 ms 3 - 10 MB NPU
DeepLabXception ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 20.845 ms 0 - 95 MB NPU
DeepLabXception ONNX float Qualcomm® QCS8450 36.87 ms 4 - 344 MB NPU
DeepLabXception ONNX float Qualcomm® Dragonwing™ IQ-9075 30.465 ms 3 - 9 MB NPU
DeepLabXception ONNX float Qualcomm® Dragonwing™ IQ-X7181 21.325 ms 85 - 85 MB NPU
DeepLabXception ONNX float Qualcomm® Dragonwing™ Q-8750 11.892 ms 2 - 241 MB NPU
DeepLabXception ONNX w8a16 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 6.33 ms 1 - 301 MB NPU
DeepLabXception ONNX w8a16 Snapdragon® 8 Elite For Galaxy Mobile 8.617 ms 1 - 294 MB NPU
DeepLabXception ONNX w8a16 Snapdragon® X2 Elite 7.989 ms 1 - 1 MB NPU
DeepLabXception ONNX w8a16 Snapdragon® X Elite 15.513 ms 45 - 45 MB NPU
DeepLabXception ONNX w8a16 Snapdragon® 8 Gen 3 Mobile 10.659 ms 2 - 398 MB NPU
DeepLabXception ONNX w8a16 Snapdragon® 8 Gen 1 Mobile 20.059 ms 2 - 413 MB NPU
DeepLabXception ONNX w8a16 Qualcomm® Dragonwing™ QCS6490 77.278 ms 1 - 5 MB NPU
DeepLabXception ONNX w8a16 Qualcomm® Dragonwing™ IQ-8275 14.78 ms 1 - 6 MB NPU
DeepLabXception ONNX w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 14.713 ms 1 - 4 MB NPU
DeepLabXception ONNX w8a16 Qualcomm® QCS8450 20.059 ms 2 - 413 MB NPU
DeepLabXception ONNX w8a16 Qualcomm® Dragonwing™ IQ-9075 15.489 ms 1 - 5 MB NPU
DeepLabXception ONNX w8a16 Qualcomm® Dragonwing™ IQ-X7181 15.513 ms 45 - 45 MB NPU
DeepLabXception ONNX w8a16 Qualcomm® Dragonwing™ Q-6690 203.181 ms 0 - 401 MB NPU
DeepLabXception ONNX w8a16 Qualcomm® Dragonwing™ Q-8750 8.617 ms 1 - 294 MB NPU
DeepLabXception ONNX w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 3.106 ms 0 - 283 MB NPU
DeepLabXception ONNX w8a8 Snapdragon® 8 Elite For Galaxy Mobile 4.159 ms 0 - 277 MB NPU
DeepLabXception ONNX w8a8 Snapdragon® X2 Elite 3.393 ms 2 - 2 MB NPU
DeepLabXception ONNX w8a8 Snapdragon® X Elite 7.741 ms 44 - 44 MB NPU
DeepLabXception ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 5.286 ms 0 - 360 MB NPU
DeepLabXception ONNX w8a8 Snapdragon® 8 Gen 1 Mobile 10.502 ms 0 - 364 MB NPU
DeepLabXception ONNX w8a8 Qualcomm® Dragonwing™ QCS6490 32.037 ms 0 - 5 MB NPU
DeepLabXception ONNX w8a8 Qualcomm® Dragonwing™ IQ-8275 7.21 ms 0 - 5 MB NPU
DeepLabXception ONNX w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 7.404 ms 1 - 4 MB NPU
DeepLabXception ONNX w8a8 Qualcomm® QCS8450 10.502 ms 0 - 364 MB NPU
DeepLabXception ONNX w8a8 Qualcomm® Dragonwing™ IQ-9075 7.171 ms 0 - 5 MB NPU
DeepLabXception ONNX w8a8 Qualcomm® Dragonwing™ IQ-X7181 7.741 ms 44 - 44 MB NPU
DeepLabXception ONNX w8a8 Qualcomm® Dragonwing™ Q-6690 107.982 ms 0 - 290 MB NPU
DeepLabXception ONNX w8a8 Qualcomm® Dragonwing™ Q-7790 11.822 ms 1 - 270 MB NPU
DeepLabXception ONNX w8a8 Qualcomm® Dragonwing™ Q-8750 4.159 ms 0 - 277 MB NPU
DeepLabXception ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 11.822 ms 1 - 270 MB NPU
DeepLabXception QNN_DLC float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 9.36 ms 3 - 247 MB NPU
DeepLabXception QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 12.13 ms 0 - 239 MB NPU
DeepLabXception QNN_DLC float Snapdragon® X2 Elite 11.18 ms 3 - 3 MB NPU
DeepLabXception QNN_DLC float Snapdragon® X Elite 22.822 ms 3 - 3 MB NPU
DeepLabXception QNN_DLC float Snapdragon® 8 Gen 3 Mobile 15.985 ms 0 - 345 MB NPU
DeepLabXception QNN_DLC float Snapdragon® 8 Gen 1 Mobile 37.207 ms 0 - 333 MB NPU
DeepLabXception QNN_DLC float Qualcomm® Dragonwing™ IQ-8275 33.662 ms 3 - 9 MB NPU
DeepLabXception QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 21.819 ms 3 - 5 MB NPU
DeepLabXception QNN_DLC float Qualcomm® SA8775P 30.354 ms 1 - 233 MB NPU
DeepLabXception QNN_DLC float Qualcomm® SA8650P 30.354 ms 1 - 233 MB NPU
DeepLabXception QNN_DLC float Qualcomm® SA8255P 30.354 ms 1 - 233 MB NPU
DeepLabXception QNN_DLC float Qualcomm® QCS8450 37.207 ms 0 - 333 MB NPU
DeepLabXception QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 30.652 ms 3 - 8 MB NPU
DeepLabXception QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 22.822 ms 3 - 3 MB NPU
DeepLabXception QNN_DLC float Qualcomm® Dragonwing™ Q-8750 12.13 ms 0 - 239 MB NPU
DeepLabXception QNN_DLC float Qualcomm® SA8295P 38.382 ms 0 - 223 MB NPU
DeepLabXception QNN_DLC w8a16 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 6.826 ms 1 - 293 MB NPU
DeepLabXception QNN_DLC w8a16 Snapdragon® 8 Elite For Galaxy Mobile 9.276 ms 1 - 286 MB NPU
DeepLabXception QNN_DLC w8a16 Snapdragon® X2 Elite 9.078 ms 1 - 1 MB NPU
DeepLabXception QNN_DLC w8a16 Snapdragon® X Elite 17.191 ms 1 - 1 MB NPU
DeepLabXception QNN_DLC w8a16 Snapdragon® 8 Gen 3 Mobile 11.812 ms 1 - 395 MB NPU
DeepLabXception QNN_DLC w8a16 Snapdragon® 8 Gen 1 Mobile 21.449 ms 1 - 408 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS6490 79.658 ms 1 - 5 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-8275 16.339 ms 1 - 6 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 16.072 ms 1 - 4 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® QCS8450 21.449 ms 1 - 408 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-9075 16.437 ms 3 - 7 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-X7181 17.191 ms 1 - 1 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-6690 234.333 ms 1 - 380 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-8750 9.276 ms 1 - 286 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® SA7255P 41.059 ms 2 - 286 MB NPU
DeepLabXception QNN_DLC w8a16 Qualcomm® SA8295P 23.435 ms 0 - 288 MB NPU
DeepLabXception QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 3.253 ms 1 - 274 MB NPU
DeepLabXception QNN_DLC w8a8 Snapdragon® 8 Elite For Galaxy Mobile 4.306 ms 1 - 268 MB NPU
DeepLabXception QNN_DLC w8a8 Snapdragon® X2 Elite 3.892 ms 1 - 1 MB NPU
DeepLabXception QNN_DLC w8a8 Snapdragon® X Elite 8.588 ms 1 - 1 MB NPU
DeepLabXception QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 5.67 ms 1 - 351 MB NPU
DeepLabXception QNN_DLC w8a8 Snapdragon® 8 Gen 1 Mobile 12.163 ms 1 - 356 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS6490 32.712 ms 1 - 4 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-8275 7.709 ms 1 - 5 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 7.882 ms 1 - 3 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® QCS8450 12.163 ms 1 - 356 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-9075 7.776 ms 2 - 5 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-X7181 8.588 ms 1 - 1 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-6690 118.974 ms 1 - 286 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-7790 11.813 ms 1 - 267 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-8750 4.306 ms 1 - 268 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® SA7255P 20.266 ms 1 - 263 MB NPU
DeepLabXception QNN_DLC w8a8 Qualcomm® SA8295P 12.566 ms 1 - 268 MB NPU
DeepLabXception QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 11.813 ms 1 - 267 MB NPU
DeepLabXception TFLITE float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 9.391 ms 0 - 349 MB NPU
DeepLabXception TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 12.222 ms 0 - 341 MB NPU
DeepLabXception TFLITE float Snapdragon® 8 Gen 3 Mobile 16.008 ms 0 - 446 MB NPU
DeepLabXception TFLITE float Snapdragon® 8 Gen 1 Mobile 37.881 ms 0 - 434 MB NPU
DeepLabXception TFLITE float Qualcomm® Dragonwing™ IQ-8275 33.779 ms 0 - 94 MB NPU
DeepLabXception TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 21.837 ms 0 - 3 MB NPU
DeepLabXception TFLITE float Qualcomm® SA8775P 30.294 ms 0 - 334 MB NPU
DeepLabXception TFLITE float Qualcomm® SA8650P 30.294 ms 0 - 334 MB NPU
DeepLabXception TFLITE float Qualcomm® SA8255P 30.294 ms 0 - 334 MB NPU
DeepLabXception TFLITE float Qualcomm® QCS8450 37.881 ms 0 - 434 MB NPU
DeepLabXception TFLITE float Qualcomm® Dragonwing™ IQ-9075 31.043 ms 0 - 93 MB NPU
DeepLabXception TFLITE float Qualcomm® Dragonwing™ Q-8750 12.222 ms 0 - 341 MB NPU
DeepLabXception TFLITE float Qualcomm® SA7255P 112.508 ms 1 - 332 MB NPU
DeepLabXception TFLITE float Qualcomm® SA8295P 38.608 ms 0 - 323 MB NPU
DeepLabXception TFLITE w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 2.899 ms 0 - 262 MB NPU
DeepLabXception TFLITE w8a8 Snapdragon® 8 Elite For Galaxy Mobile 3.862 ms 0 - 262 MB NPU
DeepLabXception TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 4.89 ms 0 - 347 MB NPU
DeepLabXception TFLITE w8a8 Snapdragon® 8 Gen 1 Mobile 10.575 ms 0 - 348 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® Dragonwing™ QCS6490 31.604 ms 0 - 58 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® Dragonwing™ IQ-8275 6.746 ms 0 - 48 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 6.869 ms 0 - 2 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® SA8775P 7.125 ms 0 - 258 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® SA8650P 7.125 ms 0 - 258 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® SA8255P 7.125 ms 0 - 258 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® QCS8450 10.575 ms 0 - 348 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® Dragonwing™ IQ-9075 6.915 ms 0 - 47 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® Dragonwing™ Q-6690 113.221 ms 0 - 273 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® Dragonwing™ Q-7790 10.807 ms 0 - 255 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® Dragonwing™ Q-8750 3.862 ms 0 - 262 MB NPU
DeepLabXception TFLITE w8a8 Qualcomm® SA8295P 11.693 ms 0 - 262 MB NPU
DeepLabXception TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 10.807 ms 0 - 255 MB NPU

License

  • The license for the original implementation of DeepLabXception can be found here.

References

Community

Downloads last month

-

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

Paper for qualcomm/DeepLabXception