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 2.45, ONNX Runtime 1.25.0	Download
ONNX	w8a8	Universal	QAIRT 2.45, ONNX Runtime 1.25.0	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a8	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 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:

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

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
DeepLabXception	ONNX	float	Snapdragon® X2 Elite	10.45 ms	3 - 3 MB	NPU
DeepLabXception	ONNX	float	Snapdragon® X Elite	22.92 ms	86 - 86 MB	NPU
DeepLabXception	ONNX	float	Snapdragon® 8 Gen 3 Mobile	17.187 ms	3 - 361 MB	NPU
DeepLabXception	ONNX	float	Snapdragon® 8 Gen 1 Mobile	50.62 ms	4 - 357 MB	NPU
DeepLabXception	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	22.867 ms	0 - 95 MB	NPU
DeepLabXception	ONNX	float	Qualcomm® QCS8450	50.62 ms	4 - 357 MB	NPU
DeepLabXception	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	31.693 ms	3 - 8 MB	NPU
DeepLabXception	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	8.983 ms	2 - 254 MB	NPU
DeepLabXception	ONNX	float	Snapdragon® 8 Elite Mobile	13.232 ms	2 - 250 MB	NPU
DeepLabXception	ONNX	float	Qualcomm® Dragonwing™ Q-8750	13.232 ms	2 - 250 MB	NPU
DeepLabXception	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	22.92 ms	86 - 86 MB	NPU
DeepLabXception	ONNX	w8a8	Snapdragon® X2 Elite	3.409 ms	2 - 2 MB	NPU
DeepLabXception	ONNX	w8a8	Snapdragon® X Elite	7.772 ms	44 - 44 MB	NPU
DeepLabXception	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	5.28 ms	0 - 385 MB	NPU
DeepLabXception	ONNX	w8a8	Snapdragon® 8 Gen 1 Mobile	12.49 ms	1 - 379 MB	NPU
DeepLabXception	ONNX	w8a8	Qualcomm® Dragonwing™ QCS6490	45.724 ms	0 - 4 MB	NPU
DeepLabXception	ONNX	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	7.455 ms	1 - 4 MB	NPU
DeepLabXception	ONNX	w8a8	Qualcomm® QCS8450	12.49 ms	1 - 379 MB	NPU
DeepLabXception	ONNX	w8a8	Snapdragon® 8 Elite Mobile	4.232 ms	0 - 289 MB	NPU
DeepLabXception	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	11.747 ms	1 - 280 MB	NPU
DeepLabXception	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	3.163 ms	0 - 298 MB	NPU
DeepLabXception	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-9075	7.334 ms	0 - 4 MB	NPU
DeepLabXception	ONNX	w8a8	Qualcomm® Dragonwing™ Q-6690	91.646 ms	1 - 309 MB	NPU
DeepLabXception	ONNX	w8a8	Qualcomm® Dragonwing™ Q-7790	11.747 ms	1 - 280 MB	NPU
DeepLabXception	ONNX	w8a8	Qualcomm® Dragonwing™ Q-8750	4.232 ms	0 - 289 MB	NPU
DeepLabXception	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-X7181	7.772 ms	44 - 44 MB	NPU
DeepLabXception	QNN_DLC	float	Snapdragon® X2 Elite	11.514 ms	3 - 3 MB	NPU
DeepLabXception	QNN_DLC	float	Snapdragon® X Elite	22.948 ms	3 - 3 MB	NPU
DeepLabXception	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	16.134 ms	0 - 352 MB	NPU
DeepLabXception	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	45.828 ms	3 - 355 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® QCS8275	112.981 ms	1 - 244 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	22.061 ms	3 - 5 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® SA8775P	30.632 ms	0 - 244 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® SA8650P	30.632 ms	0 - 244 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® SA8255P	30.632 ms	0 - 244 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® QCS8450	45.828 ms	3 - 355 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	31.45 ms	3 - 8 MB	NPU
DeepLabXception	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	9.645 ms	3 - 260 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® SA7255P	112.981 ms	1 - 244 MB	NPU
DeepLabXception	QNN_DLC	float	Snapdragon® 8 Elite Mobile	12.183 ms	0 - 250 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® SA8295P	39.12 ms	0 - 240 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	12.183 ms	0 - 250 MB	NPU
DeepLabXception	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	22.948 ms	3 - 3 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Snapdragon® X2 Elite	3.858 ms	1 - 1 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Snapdragon® X Elite	8.573 ms	1 - 1 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	5.556 ms	1 - 370 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Snapdragon® 8 Gen 1 Mobile	13.676 ms	1 - 365 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS6490	46.651 ms	3 - 5 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® QCS8275	20.339 ms	1 - 283 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	7.874 ms	1 - 3 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® SA8775P	8.076 ms	1 - 286 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® SA8650P	8.076 ms	1 - 286 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® SA8255P	8.076 ms	1 - 286 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® QCS8450	13.676 ms	1 - 365 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Snapdragon® 8 Elite Mobile	4.367 ms	1 - 289 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	12.807 ms	1 - 277 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	3.281 ms	1 - 282 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-9075	7.775 ms	3 - 5 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-6690	93.567 ms	1 - 302 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® SA7255P	20.339 ms	1 - 283 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® SA8295P	12.726 ms	0 - 279 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-7790	12.807 ms	1 - 277 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-8750	4.367 ms	1 - 289 MB	NPU
DeepLabXception	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-X7181	8.573 ms	1 - 1 MB	NPU
DeepLabXception	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	16.201 ms	0 - 455 MB	NPU
DeepLabXception	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	45.404 ms	1 - 451 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® QCS8275	113.151 ms	1 - 343 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	21.309 ms	0 - 3 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® SA8775P	30.633 ms	0 - 344 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® SA8650P	30.633 ms	0 - 344 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® SA8255P	30.633 ms	0 - 344 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® QCS8450	45.404 ms	1 - 451 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	31.078 ms	0 - 93 MB	NPU
DeepLabXception	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	9.292 ms	0 - 359 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® SA7255P	113.151 ms	1 - 343 MB	NPU
DeepLabXception	TFLITE	float	Snapdragon® 8 Elite Mobile	12.237 ms	0 - 355 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® SA8295P	38.739 ms	0 - 336 MB	NPU
DeepLabXception	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	12.237 ms	0 - 355 MB	NPU
DeepLabXception	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	4.989 ms	0 - 369 MB	NPU
DeepLabXception	TFLITE	w8a8	Snapdragon® 8 Gen 1 Mobile	12.409 ms	0 - 359 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS6490	44.859 ms	0 - 58 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® QCS8275	19.034 ms	0 - 284 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	6.952 ms	0 - 2 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® SA8775P	7.275 ms	0 - 285 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® SA8650P	7.275 ms	0 - 285 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® SA8255P	7.275 ms	0 - 285 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® QCS8450	12.409 ms	0 - 359 MB	NPU
DeepLabXception	TFLITE	w8a8	Snapdragon® 8 Elite Mobile	3.908 ms	1 - 269 MB	NPU
DeepLabXception	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	11.178 ms	0 - 265 MB	NPU
DeepLabXception	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	2.914 ms	0 - 285 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-9075	6.963 ms	0 - 47 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-6690	91.018 ms	0 - 291 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® SA7255P	19.034 ms	0 - 284 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® SA8295P	11.616 ms	0 - 276 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-7790	11.178 ms	0 - 265 MB	NPU
DeepLabXception	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-8750	3.908 ms	1 - 269 MB	NPU

License

The license for the original implementation of DeepLabXception 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/DeepLabXception

Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation

Paper • 1802.02611 • Published Feb 7, 2018