QuickSRNetSmall: Optimized for Qualcomm Devices
QuickSRNet Small is designed for upscaling images on mobile platforms to sharpen in real-time.
This is based on the implementation of QuickSRNetSmall 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 QuickSRNetSmall 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 QuickSRNetSmall on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.super_resolution
Model Stats:
- Model checkpoint: quicksrnet_small_3x_checkpoint
- Input resolution: 128x128
- Number of parameters: 33.3K
- Model size (float): 133 KB
- Model size (w8a8): 41.7 KB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| QuickSRNetSmall | ONNX | float | Snapdragon® X2 Elite | 0.363 ms | 213 - 213 MB | NPU |
| QuickSRNetSmall | ONNX | float | Snapdragon® X Elite | 0.88 ms | 149 - 149 MB | NPU |
| QuickSRNetSmall | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.544 ms | 0 - 15 MB | NPU |
| QuickSRNetSmall | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 1.222 ms | 0 - 23 MB | NPU |
| QuickSRNetSmall | ONNX | float | Qualcomm® QCS8550 (Proxy) | 0.844 ms | 0 - 15 MB | NPU |
| QuickSRNetSmall | ONNX | float | Qualcomm® QCS8450 | 1.222 ms | 0 - 23 MB | NPU |
| QuickSRNetSmall | ONNX | float | Snapdragon® 8 Elite Mobile | 0.444 ms | 0 - 22 MB | NPU |
| QuickSRNetSmall | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.392 ms | 0 - 14 MB | NPU |
| QuickSRNetSmall | ONNX | float | Qualcomm® QCS9075 | 1.102 ms | 3 - 48 MB | NPU |
| QuickSRNetSmall | ONNX | float | Qualcomm® QCS8750 | 0.444 ms | 0 - 22 MB | NPU |
| QuickSRNetSmall | ONNX | float | Qualcomm® QCS7181 | 0.88 ms | 149 - 149 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Snapdragon® X2 Elite | 0.195 ms | 181 - 181 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Snapdragon® X Elite | 0.494 ms | 149 - 149 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.318 ms | 0 - 16 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.571 ms | 0 - 25 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Qualcomm® QCS6490 | 1.625 ms | 0 - 45 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.502 ms | 0 - 15 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Qualcomm® QCS8450 | 0.571 ms | 0 - 25 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.208 ms | 0 - 15 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.525 ms | 0 - 14 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Qualcomm® QCM6690 | 1.795 ms | 0 - 128 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Qualcomm® QCS9075 | 0.65 ms | 0 - 45 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Snapdragon® 8 Elite Mobile | 0.259 ms | 0 - 22 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Qualcomm® QCS7790 | 0.525 ms | 0 - 14 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Qualcomm® QCS8750 | 0.259 ms | 0 - 22 MB | NPU |
| QuickSRNetSmall | ONNX | w8a8 | Qualcomm® QCS7181 | 0.494 ms | 149 - 149 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Snapdragon® X2 Elite | 0.406 ms | 0 - 0 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Snapdragon® X Elite | 0.822 ms | 0 - 0 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 0.447 ms | 0 - 27 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 1.035 ms | 0 - 30 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Qualcomm® QCS8275 | 1.879 ms | 0 - 20 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 0.737 ms | 0 - 11 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Qualcomm® QCS8450 | 1.035 ms | 0 - 30 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 0.352 ms | 0 - 20 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Qualcomm® SA8295P | 1.411 ms | 0 - 18 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.325 ms | 0 - 21 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Qualcomm® SA7255P | 1.879 ms | 0 - 20 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Qualcomm® QCS9075 | 1.089 ms | 0 - 5 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Qualcomm® QCS8750 | 0.352 ms | 0 - 20 MB | NPU |
| QuickSRNetSmall | QNN_DLC | float | Qualcomm® QCS7181 | 0.822 ms | 0 - 0 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 0.26 ms | 0 - 0 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Snapdragon® X Elite | 0.456 ms | 0 - 0 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.22 ms | 0 - 27 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.418 ms | 0 - 31 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® QCS6490 | 1.271 ms | 0 - 2 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® QCS8275 | 0.821 ms | 0 - 19 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.365 ms | 0 - 14 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® QCS8450 | 0.418 ms | 0 - 31 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.163 ms | 0 - 22 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.359 ms | 0 - 18 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® QCM6690 | 1.28 ms | 0 - 17 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 0.495 ms | 2 - 4 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® SA7255P | 0.821 ms | 0 - 19 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Snapdragon® 8 Elite Mobile | 0.193 ms | 0 - 24 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® SA8295P | 0.716 ms | 0 - 17 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® QCS7790 | 0.359 ms | 0 - 18 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® QCS8750 | 0.193 ms | 0 - 24 MB | NPU |
| QuickSRNetSmall | QNN_DLC | w8a8 | Qualcomm® QCS7181 | 0.456 ms | 0 - 0 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 0.612 ms | 0 - 28 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 1.324 ms | 0 - 33 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® QCS8275 | 2.401 ms | 0 - 21 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 0.998 ms | 0 - 5 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® SA8775P | 3.501 ms | 0 - 25 MB | GPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® SA8650P | 3.501 ms | 0 - 25 MB | GPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® SA8255P | 3.501 ms | 0 - 25 MB | GPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® QCS8450 | 1.324 ms | 0 - 33 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Snapdragon® 8 Elite Mobile | 0.447 ms | 0 - 20 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® SA8295P | 1.63 ms | 0 - 17 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.393 ms | 0 - 21 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® SA7255P | 2.401 ms | 0 - 21 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® QCS9075 | 1.269 ms | 3 - 8 MB | NPU |
| QuickSRNetSmall | TFLITE | float | Qualcomm® QCS8750 | 0.447 ms | 0 - 20 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.269 ms | 0 - 27 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.467 ms | 0 - 32 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® QCS6490 | 1.195 ms | 0 - 2 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® QCS8275 | 0.902 ms | 0 - 19 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.413 ms | 0 - 7 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® SA8775P | 4.046 ms | 4 - 30 MB | GPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® SA8650P | 4.046 ms | 4 - 30 MB | GPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® SA8255P | 4.046 ms | 4 - 30 MB | GPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® QCS8450 | 0.467 ms | 0 - 32 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.177 ms | 0 - 21 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.423 ms | 0 - 17 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® QCM6690 | 1.495 ms | 0 - 17 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® QCS9075 | 0.52 ms | 0 - 3 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® SA7255P | 0.902 ms | 0 - 19 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Snapdragon® 8 Elite Mobile | 0.22 ms | 0 - 19 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® SA8295P | 0.813 ms | 0 - 17 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® QCS7790 | 0.423 ms | 0 - 17 MB | NPU |
| QuickSRNetSmall | TFLITE | w8a8 | Qualcomm® QCS8750 | 0.22 ms | 0 - 19 MB | NPU |
License
- The license for the original implementation of QuickSRNetSmall 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.
