Segformer-Base: Optimized for Qualcomm Devices
Segformer Base is a machine learning model that predicts masks and classes of objects in an image.
This is based on the implementation of Segformer-Base 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 Segformer-Base 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 Segformer-Base on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.semantic_segmentation
Model Stats:
- Input resolution: 512x512
- Model checkpoint: nvidia/segformer-b0-finetuned-ade-512-512
- Model size (float): 14.4 MB
- Model size (w8a16): 4.57 MB
- Model size (w8a8): 3.90 MB
- Number of output classes: 150
- Number of parameters: 3.75M
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| Segformer-Base | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 11.308 ms | 21 - 203 MB | NPU |
| Segformer-Base | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 12.442 ms | 23 - 202 MB | NPU |
| Segformer-Base | ONNX | float | Snapdragon® X2 Elite | 11.389 ms | 35 - 35 MB | NPU |
| Segformer-Base | ONNX | float | Snapdragon® X Elite | 21.038 ms | 35 - 35 MB | NPU |
| Segformer-Base | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 15.651 ms | 27 - 243 MB | NPU |
| Segformer-Base | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 23.055 ms | 1 - 220 MB | NPU |
| Segformer-Base | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 21.639 ms | 22 - 26 MB | NPU |
| Segformer-Base | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 20.613 ms | 19 - 26 MB | NPU |
| Segformer-Base | ONNX | float | Qualcomm® QCS8450 | 23.055 ms | 1 - 220 MB | NPU |
| Segformer-Base | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 23.616 ms | 22 - 26 MB | NPU |
| Segformer-Base | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 21.038 ms | 35 - 35 MB | NPU |
| Segformer-Base | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 12.442 ms | 23 - 202 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 5.224 ms | 14 - 214 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 6.677 ms | 13 - 211 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Snapdragon® X2 Elite | 5.209 ms | 18 - 18 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Snapdragon® X Elite | 12.98 ms | 17 - 17 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 8.944 ms | 12 - 234 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 14.877 ms | 11 - 239 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 33.724 ms | 13 - 17 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 12.606 ms | 11 - 15 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 12.721 ms | 9 - 15 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Qualcomm® QCS8450 | 14.877 ms | 11 - 239 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 12.804 ms | 14 - 17 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 12.98 ms | 17 - 17 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 67.876 ms | 14 - 235 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 16.822 ms | 13 - 224 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 6.677 ms | 13 - 211 MB | NPU |
| Segformer-Base | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 16.822 ms | 13 - 224 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 1.928 ms | 0 - 193 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 2.419 ms | 0 - 183 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Snapdragon® X2 Elite | 1.965 ms | 9 - 9 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Snapdragon® X Elite | 4.992 ms | 5 - 5 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 3.242 ms | 0 - 211 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Snapdragon® 8 Gen 1 Mobile | 5.786 ms | 6 - 220 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 15.898 ms | 7 - 10 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-8275 | 4.643 ms | 5 - 10 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 4.752 ms | 5 - 11 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Qualcomm® QCS8450 | 5.786 ms | 6 - 220 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 5.221 ms | 7 - 10 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 4.992 ms | 5 - 5 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 29.698 ms | 0 - 196 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 5.369 ms | 9 - 202 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 2.419 ms | 0 - 183 MB | NPU |
| Segformer-Base | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 5.369 ms | 9 - 202 MB | NPU |
| Segformer-Base | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 11.632 ms | 3 - 187 MB | NPU |
| Segformer-Base | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 13.0 ms | 3 - 183 MB | NPU |
| Segformer-Base | QNN_DLC | float | Snapdragon® X2 Elite | 12.155 ms | 3 - 3 MB | NPU |
| Segformer-Base | QNN_DLC | float | Snapdragon® X Elite | 23.717 ms | 3 - 3 MB | NPU |
| Segformer-Base | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 16.296 ms | 3 - 213 MB | NPU |
| Segformer-Base | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 27.02 ms | 3 - 214 MB | NPU |
| Segformer-Base | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 23.29 ms | 3 - 18 MB | NPU |
| Segformer-Base | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 22.633 ms | 3 - 5 MB | NPU |
| Segformer-Base | QNN_DLC | float | Qualcomm® SA8650P | 24.435 ms | 0 - 193 MB | NPU |
| Segformer-Base | QNN_DLC | float | Qualcomm® SA8255P | 24.435 ms | 0 - 193 MB | NPU |
| Segformer-Base | QNN_DLC | float | Qualcomm® QCS8450 | 27.02 ms | 3 - 214 MB | NPU |
| Segformer-Base | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 25.685 ms | 3 - 17 MB | NPU |
| Segformer-Base | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 23.717 ms | 3 - 3 MB | NPU |
| Segformer-Base | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 13.0 ms | 3 - 183 MB | NPU |
| Segformer-Base | QNN_DLC | float | Qualcomm® SA8295P | 30.307 ms | 0 - 176 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 5.69 ms | 2 - 188 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 7.554 ms | 2 - 186 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 6.081 ms | 2 - 2 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Snapdragon® X Elite | 15.119 ms | 2 - 2 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 10.519 ms | 2 - 212 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 14.185 ms | 1 - 10 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 14.7 ms | 2 - 3 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Qualcomm® SA8650P | 20.163 ms | 2 - 206 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Qualcomm® SA8255P | 20.163 ms | 2 - 206 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 15.704 ms | 4 - 12 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 15.119 ms | 2 - 2 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 100.509 ms | 2 - 214 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 17.31 ms | 2 - 198 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 7.554 ms | 2 - 186 MB | NPU |
| Segformer-Base | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 17.31 ms | 2 - 198 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.281 ms | 1 - 177 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 2.861 ms | 1 - 171 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 2.629 ms | 1 - 1 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Snapdragon® X Elite | 6.377 ms | 1 - 1 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 3.983 ms | 0 - 192 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Snapdragon® 8 Gen 1 Mobile | 6.833 ms | 0 - 197 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 17.213 ms | 3 - 8 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-8275 | 5.514 ms | 1 - 7 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.823 ms | 1 - 3 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® SA8650P | 6.393 ms | 1 - 174 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® SA8255P | 6.393 ms | 1 - 174 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® QCS8450 | 6.833 ms | 0 - 197 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 6.206 ms | 1 - 6 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 6.377 ms | 1 - 1 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 32.199 ms | 1 - 184 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 6.264 ms | 0 - 174 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 2.861 ms | 1 - 171 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® SA7255P | 10.82 ms | 1 - 172 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Qualcomm® SA8295P | 7.661 ms | 0 - 175 MB | NPU |
| Segformer-Base | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 6.264 ms | 0 - 174 MB | NPU |
| Segformer-Base | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 11.623 ms | 8 - 187 MB | NPU |
| Segformer-Base | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 13.053 ms | 9 - 192 MB | NPU |
| Segformer-Base | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 17.502 ms | 9 - 214 MB | NPU |
| Segformer-Base | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 26.952 ms | 10 - 220 MB | NPU |
| Segformer-Base | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 23.02 ms | 9 - 33 MB | NPU |
| Segformer-Base | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 22.581 ms | 9 - 21 MB | NPU |
| Segformer-Base | TFLITE | float | Qualcomm® SA8650P | 24.464 ms | 9 - 202 MB | NPU |
| Segformer-Base | TFLITE | float | Qualcomm® SA8255P | 24.464 ms | 9 - 202 MB | NPU |
| Segformer-Base | TFLITE | float | Qualcomm® QCS8450 | 26.952 ms | 10 - 220 MB | NPU |
| Segformer-Base | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 25.502 ms | 8 - 31 MB | NPU |
| Segformer-Base | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 13.053 ms | 9 - 192 MB | NPU |
| Segformer-Base | TFLITE | float | Qualcomm® SA8295P | 30.399 ms | 9 - 181 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 4.823 ms | 1 - 181 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 5.86 ms | 2 - 187 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 7.391 ms | 2 - 200 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 11.817 ms | 2 - 206 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 137.47 ms | 13 - 49 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-8275 | 10.0 ms | 2 - 13 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 10.674 ms | 2 - 6 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® SA8650P | 10.908 ms | 2 - 183 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® SA8255P | 10.908 ms | 2 - 183 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® QCS8450 | 11.817 ms | 2 - 206 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 11.225 ms | 2 - 13 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 95.802 ms | 15 - 185 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 38.936 ms | 13 - 75 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 5.86 ms | 2 - 187 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Qualcomm® SA8295P | 13.349 ms | 2 - 185 MB | NPU |
| Segformer-Base | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 38.936 ms | 13 - 75 MB | NPU |
License
- The license for the original implementation of Segformer-Base can be found here.
References
- SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers
- Source Model Implementation
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.
