NAFSSR: Stereo Image Super-Resolution Using NAFNet
Paper • 2204.08714 • Published
NAFSSR: Optimized for Qualcomm Devices
NAFNET is designed for lightweight real-time upscaling of images.
This is based on the implementation of NAFSSR 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 |
|---|---|---|---|---|
| PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X2 Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X2 Elite | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X Elite | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8775P | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-6690 | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA7255P | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 7 Gen 4 Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Mobile | QAIRT 2.45 | Download |
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45 | Download |
For more device-specific assets and performance metrics, visit NAFSSR 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 NAFSSR on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.super_resolution
Model Stats:
- Model checkpoint: NAFSSR-S_2x
- Input resolution: 128x128
- Number of parameters: 1.54M
- Model size (float): 6.2 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 79.921 ms | 8 - 8 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 199.205 ms | 41 - 41 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 128.785 ms | 4 - 11 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | 234.63 ms | 4 - 18 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 188.206 ms | 0 - 41 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8450 | 234.63 ms | 4 - 18 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 193.903 ms | 3 - 6 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 199.205 ms | 41 - 41 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 99.61 ms | 5 - 12 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Mobile | 99.61 ms | 5 - 12 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 78.386 ms | 7 - 14 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X2 Elite | 64.466 ms | 3 - 3 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X Elite | 190.116 ms | 35 - 35 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 118.034 ms | 2 - 9 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 180.689 ms | 0 - 42 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 225.751 ms | 2 - 4 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 190.116 ms | 35 - 35 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 1590.304 ms | 0 - 7 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 204.318 ms | 1 - 7 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 85.856 ms | 2 - 9 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 85.856 ms | 2 - 9 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 61.718 ms | 2 - 9 MB | NPU |
| NAFSSR | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 204.318 ms | 1 - 7 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 79.67 ms | 0 - 0 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 200.065 ms | 0 - 0 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 127.268 ms | 1 - 7 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | 234.696 ms | 1 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8275 | 355.399 ms | 1 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 189.946 ms | 1 - 4 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | 195.306 ms | 1 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® SA8650P | 195.306 ms | 1 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® SA8255P | 195.306 ms | 1 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 | 234.696 ms | 1 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | 274.916 ms | 0 - 4 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-X7181 | 200.065 ms | 0 - 0 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ Q-8750 | 100.708 ms | 1 - 13 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | 355.399 ms | 1 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 228.169 ms | 1 - 6 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | 100.708 ms | 1 - 13 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | 77.608 ms | 0 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X2 Elite | 64.808 ms | 0 - 0 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X Elite | 190.324 ms | 0 - 0 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Gen 3 Mobile | 118.127 ms | 0 - 7 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8275 | 299.2 ms | 0 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 180.534 ms | 0 - 2 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8775P | 184.654 ms | 0 - 9 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8650P | 184.654 ms | 0 - 9 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8255P | 184.654 ms | 0 - 9 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 182.842 ms | 0 - 3 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 190.324 ms | 0 - 0 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 1576.863 ms | 0 - 7 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 205.122 ms | 0 - 7 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 86.34 ms | 2 - 14 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA7255P | 299.2 ms | 0 - 10 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Mobile | 86.34 ms | 2 - 14 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 61.76 ms | 0 - 9 MB | NPU |
| NAFSSR | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 7 Gen 4 Mobile | 205.122 ms | 0 - 7 MB | NPU |
License
- The license for the original implementation of NAFSSR 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.