Rethinking Atrous Convolution for Semantic Image Segmentation
Paper
• 1706.05587 • Published
DeepLabV3-Plus-MobileNet: Optimized for Qualcomm Devices
DeepLabV3 is designed for semantic segmentation at multiple scales, trained on the various datasets. It uses MobileNet as a backbone.
This is based on the implementation of DeepLabV3-Plus-MobileNet 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 1.24.1 | Download |
| ONNX | w8a16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| ONNX | w8a8 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | Download |
| QNN_DLC | float | Universal | QAIRT 2.43 | Download |
| QNN_DLC | w8a16 | Universal | QAIRT 2.43 | Download |
| QNN_DLC | w8a8 | Universal | QAIRT 2.43 | Download |
| TFLITE | float | Universal | QAIRT 2.43, TFLite 2.17.0 | Download |
| TFLITE | w8a8 | Universal | QAIRT 2.43, TFLite 2.17.0 | Download |
For more device-specific assets and performance metrics, visit DeepLabV3-Plus-MobileNet 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 DeepLabV3-Plus-MobileNet on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.semantic_segmentation
Model Stats:
- Model checkpoint: VOC2012
- Input resolution: 513x513
- Number of output classes: 21
- Number of parameters: 5.80M
- Model size (float): 22.2 MB
- Model size (w8a16): 6.67 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| DeepLabV3-Plus-MobileNet | ONNX | float | Snapdragon® X Elite | 11.046 ms | 10 - 10 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 7.181 ms | 4 - 210 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | float | Qualcomm® QCS8550 (Proxy) | 10.342 ms | 3 - 5 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | float | Qualcomm® QCS9075 | 18.041 ms | 3 - 6 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 5.891 ms | 1 - 169 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 4.299 ms | 4 - 177 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | float | Snapdragon® X2 Elite | 5.38 ms | 10 - 10 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Snapdragon® X Elite | 7.682 ms | 5 - 5 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 5.097 ms | 2 - 223 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Qualcomm® QCS6490 | 1221.121 ms | 89 - 92 MB | CPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 7.059 ms | 2 - 4 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Qualcomm® QCS9075 | 8.623 ms | 2 - 4 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Qualcomm® QCM6690 | 590.767 ms | 117 - 125 MB | CPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 3.947 ms | 0 - 181 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 584.353 ms | 92 - 100 MB | CPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.959 ms | 0 - 207 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a16 | Snapdragon® X2 Elite | 3.71 ms | 7 - 7 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a8 | Snapdragon® X Elite | 3.923 ms | 5 - 5 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 2.37 ms | 0 - 199 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | 3.534 ms | 0 - 15 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a8 | Qualcomm® QCS9075 | 4.461 ms | 1 - 4 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 1.899 ms | 0 - 176 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 1.463 ms | 0 - 178 MB | NPU |
| DeepLabV3-Plus-MobileNet | ONNX | w8a8 | Snapdragon® X2 Elite | 1.666 ms | 7 - 7 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Snapdragon® X Elite | 12.254 ms | 3 - 3 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 8.06 ms | 0 - 211 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 58.586 ms | 1 - 167 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 11.716 ms | 3 - 5 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Qualcomm® SA8775P | 17.522 ms | 1 - 171 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Qualcomm® QCS9075 | 20.031 ms | 3 - 8 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 19.056 ms | 3 - 213 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Qualcomm® SA7255P | 58.586 ms | 1 - 167 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Qualcomm® SA8295P | 19.058 ms | 0 - 169 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 6.253 ms | 3 - 200 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 4.295 ms | 3 - 195 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | float | Snapdragon® X2 Elite | 5.541 ms | 3 - 3 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Snapdragon® X Elite | 9.188 ms | 2 - 2 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 6.122 ms | 2 - 223 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Qualcomm® QCS6490 | 33.513 ms | 0 - 3 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Qualcomm® QCS8275 (Proxy) | 21.922 ms | 2 - 182 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 8.45 ms | 2 - 3 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Qualcomm® SA8775P | 40.073 ms | 2 - 182 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 9.794 ms | 3 - 6 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 93.114 ms | 2 - 231 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Qualcomm® QCS8450 (Proxy) | 12.268 ms | 2 - 221 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Qualcomm® SA7255P | 21.922 ms | 2 - 182 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Qualcomm® SA8295P | 13.725 ms | 2 - 197 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 4.322 ms | 2 - 180 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 12.092 ms | 2 - 181 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 3.073 ms | 1 - 196 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 4.144 ms | 2 - 2 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Snapdragon® X Elite | 4.55 ms | 1 - 1 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 2.902 ms | 1 - 197 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Qualcomm® QCS6490 | 15.45 ms | 3 - 5 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Qualcomm® QCS8275 (Proxy) | 10.794 ms | 1 - 169 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 4.125 ms | 1 - 2 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Qualcomm® SA8775P | 20.173 ms | 1 - 169 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 5.018 ms | 3 - 5 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Qualcomm® QCM6690 | 47.977 ms | 1 - 202 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Qualcomm® QCS8450 (Proxy) | 6.357 ms | 1 - 198 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Qualcomm® SA7255P | 10.794 ms | 1 - 169 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Qualcomm® SA8295P | 6.407 ms | 1 - 167 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 2.143 ms | 0 - 175 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 6.131 ms | 1 - 175 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 1.504 ms | 1 - 175 MB | NPU |
| DeepLabV3-Plus-MobileNet | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 2.054 ms | 1 - 1 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 8.061 ms | 0 - 219 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 58.594 ms | 0 - 171 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 11.47 ms | 0 - 3 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Qualcomm® SA8775P | 17.575 ms | 0 - 174 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Qualcomm® QCS9075 | 19.647 ms | 0 - 18 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 19.011 ms | 0 - 215 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Qualcomm® SA7255P | 58.594 ms | 0 - 171 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Qualcomm® SA8295P | 19.1 ms | 0 - 170 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 6.249 ms | 0 - 199 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 4.293 ms | 0 - 194 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 2.955 ms | 0 - 207 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Qualcomm® QCS6490 | 15.939 ms | 0 - 11 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Qualcomm® QCS8275 (Proxy) | 10.844 ms | 0 - 173 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 4.318 ms | 0 - 119 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Qualcomm® SA8775P | 20.29 ms | 0 - 173 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Qualcomm® QCS9075 | 5.158 ms | 0 - 10 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Qualcomm® QCM6690 | 53.545 ms | 0 - 193 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Qualcomm® QCS8450 (Proxy) | 5.653 ms | 0 - 205 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Qualcomm® SA7255P | 10.844 ms | 0 - 173 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Qualcomm® SA8295P | 6.224 ms | 0 - 172 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 2.106 ms | 0 - 177 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 6.358 ms | 0 - 177 MB | NPU |
| DeepLabV3-Plus-MobileNet | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 1.527 ms | 0 - 182 MB | NPU |
License
- The license for the original implementation of DeepLabV3-Plus-MobileNet 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.