QuickSRNetLarge: Optimized for Qualcomm Devices
QuickSRNet Large is designed for upscaling images on mobile platforms to sharpen in real-time.
This is based on the implementation of QuickSRNetLarge 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 QuickSRNetLarge 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 QuickSRNetLarge on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.super_resolution
Model Stats:
- Model checkpoint: quicksrnet_large_3x_checkpoint
- Input resolution: 128x128
- Number of parameters: 436K
- Model size (float): 1.67 MB
- Model size (w8a8): 462 KB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| QuickSRNetLarge | ONNX | float | Snapdragon® X2 Elite | 1.197 ms | 213 - 213 MB | NPU |
| QuickSRNetLarge | ONNX | float | Snapdragon® X Elite | 1.981 ms | 150 - 150 MB | NPU |
| QuickSRNetLarge | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 1.433 ms | 0 - 26 MB | NPU |
| QuickSRNetLarge | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 3.46 ms | 0 - 36 MB | NPU |
| QuickSRNetLarge | ONNX | float | Qualcomm® QCS8550 (Proxy) | 1.904 ms | 0 - 5 MB | NPU |
| QuickSRNetLarge | ONNX | float | Qualcomm® QCS8450 | 3.46 ms | 0 - 36 MB | NPU |
| QuickSRNetLarge | ONNX | float | Snapdragon® 8 Elite Mobile | 1.2 ms | 0 - 20 MB | NPU |
| QuickSRNetLarge | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.916 ms | 0 - 17 MB | NPU |
| QuickSRNetLarge | ONNX | float | Qualcomm® QCS9075 | 3.523 ms | 7 - 53 MB | NPU |
| QuickSRNetLarge | ONNX | float | Qualcomm® QCS8750 | 1.2 ms | 0 - 20 MB | NPU |
| QuickSRNetLarge | ONNX | float | Qualcomm® QCS7181 | 1.981 ms | 150 - 150 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Snapdragon® X2 Elite | 0.354 ms | 213 - 213 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Snapdragon® X Elite | 0.779 ms | 181 - 181 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.514 ms | 0 - 29 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Snapdragon® 8 Gen 1 Mobile | 1.068 ms | 2 - 35 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Qualcomm® QCS6490 | 3.063 ms | 1 - 47 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.78 ms | 0 - 6 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Qualcomm® QCS8450 | 1.068 ms | 2 - 35 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.316 ms | 0 - 21 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 1.107 ms | 0 - 21 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Qualcomm® QCM6690 | 9.06 ms | 1 - 140 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Qualcomm® QCS9075 | 0.991 ms | 0 - 46 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Snapdragon® 8 Elite Mobile | 0.383 ms | 0 - 24 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Qualcomm® QCS7790 | 1.107 ms | 0 - 21 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Qualcomm® QCS8750 | 0.383 ms | 0 - 24 MB | NPU |
| QuickSRNetLarge | ONNX | w8a8 | Qualcomm® QCS7181 | 0.779 ms | 181 - 181 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Snapdragon® X2 Elite | 1.245 ms | 0 - 0 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Snapdragon® X Elite | 1.971 ms | 0 - 0 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 1.29 ms | 0 - 37 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 3.265 ms | 0 - 43 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Qualcomm® QCS8275 | 11.87 ms | 0 - 23 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 1.805 ms | 0 - 4 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Qualcomm® QCS8450 | 3.265 ms | 0 - 43 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 1.041 ms | 0 - 23 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Qualcomm® SA8295P | 3.915 ms | 0 - 20 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.837 ms | 0 - 24 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Qualcomm® SA7255P | 11.87 ms | 0 - 23 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Qualcomm® QCS9075 | 3.363 ms | 0 - 5 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Qualcomm® QCS8750 | 1.041 ms | 0 - 23 MB | NPU |
| QuickSRNetLarge | QNN_DLC | float | Qualcomm® QCS7181 | 1.971 ms | 0 - 0 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 0.423 ms | 0 - 0 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Snapdragon® X Elite | 0.777 ms | 0 - 0 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.442 ms | 0 - 34 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.823 ms | 0 - 35 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® QCS6490 | 2.736 ms | 0 - 2 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® QCS8275 | 2.121 ms | 0 - 24 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.649 ms | 0 - 30 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® QCS8450 | 0.823 ms | 0 - 35 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.267 ms | 0 - 24 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.912 ms | 0 - 22 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® QCM6690 | 8.322 ms | 0 - 136 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 0.851 ms | 2 - 4 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® SA7255P | 2.121 ms | 0 - 24 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Snapdragon® 8 Elite Mobile | 0.374 ms | 0 - 26 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® SA8295P | 1.374 ms | 0 - 21 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® QCS7790 | 0.912 ms | 0 - 22 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® QCS8750 | 0.374 ms | 0 - 26 MB | NPU |
| QuickSRNetLarge | QNN_DLC | w8a8 | Qualcomm® QCS7181 | 0.777 ms | 0 - 0 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 1.448 ms | 0 - 39 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 3.499 ms | 0 - 44 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® QCS8275 | 12.183 ms | 3 - 28 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 2.138 ms | 0 - 2 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® SA8775P | 27.916 ms | 3 - 20 MB | GPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® SA8650P | 27.916 ms | 3 - 20 MB | GPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® SA8255P | 27.916 ms | 3 - 20 MB | GPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® QCS8450 | 3.499 ms | 0 - 44 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Snapdragon® 8 Elite Mobile | 1.081 ms | 0 - 26 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® SA8295P | 4.182 ms | 0 - 22 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.866 ms | 0 - 28 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® SA7255P | 12.183 ms | 3 - 28 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® QCS9075 | 3.704 ms | 3 - 9 MB | NPU |
| QuickSRNetLarge | TFLITE | float | Qualcomm® QCS8750 | 1.081 ms | 0 - 26 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.504 ms | 0 - 34 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.91 ms | 0 - 40 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® QCS6490 | 2.951 ms | 0 - 3 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® QCS8275 | 2.188 ms | 0 - 26 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.784 ms | 0 - 1 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® SA8775P | 28.64 ms | 4 - 26 MB | GPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® SA8650P | 28.64 ms | 4 - 26 MB | GPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® SA8255P | 28.64 ms | 4 - 26 MB | GPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® QCS8450 | 0.91 ms | 0 - 40 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.287 ms | 0 - 26 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.969 ms | 0 - 24 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® QCM6690 | 8.555 ms | 0 - 137 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® QCS9075 | 0.978 ms | 0 - 3 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® SA7255P | 2.188 ms | 0 - 26 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Snapdragon® 8 Elite Mobile | 0.386 ms | 0 - 29 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® SA8295P | 1.47 ms | 0 - 23 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® QCS7790 | 0.969 ms | 0 - 24 MB | NPU |
| QuickSRNetLarge | TFLITE | w8a8 | Qualcomm® QCS8750 | 0.386 ms | 0 - 29 MB | NPU |
License
- The license for the original implementation of QuickSRNetLarge can be found here.
References
- QuickSRNet: Plain Single-Image Super-Resolution Architecture for Faster Inference on Mobile Platforms
- 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.
