SESR-M5: Optimized for Qualcomm Devices
SESR M5 performs efficient on-device upscaling of images.
This is based on the implementation of SESR-M5 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 SESR-M5 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 SESR-M5 on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.super_resolution
Model Stats:
- Model checkpoint: sesr_m5_3x_checkpoint
- Input resolution: 128x128
- Number of parameters: 343K
- Model size (float): 1.32 MB
- Model size (w8a8): 395 KB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| SESR-M5 | ONNX | float | Snapdragon® X2 Elite | 1.139 ms | 213 - 213 MB | NPU |
| SESR-M5 | ONNX | float | Snapdragon® X Elite | 2.009 ms | 180 - 180 MB | NPU |
| SESR-M5 | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 1.377 ms | 0 - 24 MB | NPU |
| SESR-M5 | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 3.03 ms | 0 - 32 MB | NPU |
| SESR-M5 | ONNX | float | Qualcomm® QCS8550 (Proxy) | 1.938 ms | 0 - 27 MB | NPU |
| SESR-M5 | ONNX | float | Qualcomm® QCS8450 | 3.03 ms | 0 - 32 MB | NPU |
| SESR-M5 | ONNX | float | Snapdragon® 8 Elite Mobile | 1.161 ms | 0 - 22 MB | NPU |
| SESR-M5 | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.952 ms | 0 - 19 MB | NPU |
| SESR-M5 | ONNX | float | Qualcomm® QCS9075 | 3.158 ms | 7 - 53 MB | NPU |
| SESR-M5 | ONNX | float | Qualcomm® QCS8750 | 1.161 ms | 0 - 22 MB | NPU |
| SESR-M5 | ONNX | float | Qualcomm® QCS7181 | 2.009 ms | 180 - 180 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Snapdragon® X2 Elite | 0.394 ms | 213 - 213 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Snapdragon® X Elite | 0.791 ms | 150 - 150 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.53 ms | 0 - 22 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Snapdragon® 8 Gen 1 Mobile | 1.032 ms | 1 - 32 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Qualcomm® QCS6490 | 2.909 ms | 0 - 46 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.779 ms | 0 - 3 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Qualcomm® QCS8450 | 1.032 ms | 1 - 32 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.374 ms | 0 - 21 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.963 ms | 0 - 133 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Qualcomm® QCM6690 | 7.416 ms | 1 - 138 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Qualcomm® QCS9075 | 0.953 ms | 0 - 46 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Snapdragon® 8 Elite Mobile | 0.425 ms | 0 - 24 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Qualcomm® QCS7790 | 0.963 ms | 0 - 133 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Qualcomm® QCS8750 | 0.425 ms | 0 - 24 MB | NPU |
| SESR-M5 | ONNX | w8a8 | Qualcomm® QCS7181 | 0.791 ms | 150 - 150 MB | NPU |
| SESR-M5 | QNN_DLC | float | Snapdragon® X2 Elite | 1.182 ms | 0 - 0 MB | NPU |
| SESR-M5 | QNN_DLC | float | Snapdragon® X Elite | 2.05 ms | 0 - 0 MB | NPU |
| SESR-M5 | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 1.274 ms | 0 - 35 MB | NPU |
| SESR-M5 | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 2.792 ms | 0 - 41 MB | NPU |
| SESR-M5 | QNN_DLC | float | Qualcomm® QCS8275 | 10.525 ms | 0 - 23 MB | NPU |
| SESR-M5 | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 1.866 ms | 0 - 69 MB | NPU |
| SESR-M5 | QNN_DLC | float | Qualcomm® QCS8450 | 2.792 ms | 0 - 41 MB | NPU |
| SESR-M5 | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 1.031 ms | 0 - 24 MB | NPU |
| SESR-M5 | QNN_DLC | float | Qualcomm® SA8295P | 3.478 ms | 0 - 21 MB | NPU |
| SESR-M5 | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.873 ms | 0 - 25 MB | NPU |
| SESR-M5 | QNN_DLC | float | Qualcomm® SA7255P | 10.525 ms | 0 - 23 MB | NPU |
| SESR-M5 | QNN_DLC | float | Qualcomm® QCS9075 | 3.056 ms | 0 - 5 MB | NPU |
| SESR-M5 | QNN_DLC | float | Qualcomm® QCS8750 | 1.031 ms | 0 - 24 MB | NPU |
| SESR-M5 | QNN_DLC | float | Qualcomm® QCS7181 | 2.05 ms | 0 - 0 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 0.43 ms | 0 - 0 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Snapdragon® X Elite | 0.715 ms | 0 - 0 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.439 ms | 0 - 32 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.85 ms | 0 - 38 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® QCS6490 | 2.658 ms | 0 - 2 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® QCS8275 | 1.96 ms | 0 - 24 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.6 ms | 0 - 63 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® QCS8450 | 0.85 ms | 0 - 38 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.299 ms | 0 - 25 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.788 ms | 0 - 22 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® QCM6690 | 6.812 ms | 0 - 134 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 0.896 ms | 0 - 2 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® SA7255P | 1.96 ms | 0 - 24 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Snapdragon® 8 Elite Mobile | 0.38 ms | 0 - 28 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® SA8295P | 1.331 ms | 0 - 22 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® QCS7790 | 0.788 ms | 0 - 22 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® QCS8750 | 0.38 ms | 0 - 28 MB | NPU |
| SESR-M5 | QNN_DLC | w8a8 | Qualcomm® QCS7181 | 0.715 ms | 0 - 0 MB | NPU |
| SESR-M5 | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 1.28 ms | 0 - 35 MB | NPU |
| SESR-M5 | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 2.805 ms | 0 - 38 MB | NPU |
| SESR-M5 | TFLITE | float | Qualcomm® QCS8275 | 10.581 ms | 3 - 27 MB | NPU |
| SESR-M5 | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 1.882 ms | 0 - 2 MB | NPU |
| SESR-M5 | TFLITE | float | Qualcomm® SA8775P | 21.03 ms | 3 - 20 MB | GPU |
| SESR-M5 | TFLITE | float | Qualcomm® SA8650P | 21.03 ms | 3 - 20 MB | GPU |
| SESR-M5 | TFLITE | float | Qualcomm® SA8255P | 21.03 ms | 3 - 20 MB | GPU |
| SESR-M5 | TFLITE | float | Qualcomm® QCS8450 | 2.805 ms | 0 - 38 MB | NPU |
| SESR-M5 | TFLITE | float | Snapdragon® 8 Elite Mobile | 1.03 ms | 0 - 24 MB | NPU |
| SESR-M5 | TFLITE | float | Qualcomm® SA8295P | 3.493 ms | 0 - 22 MB | NPU |
| SESR-M5 | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.875 ms | 0 - 26 MB | NPU |
| SESR-M5 | TFLITE | float | Qualcomm® SA7255P | 10.581 ms | 3 - 27 MB | NPU |
| SESR-M5 | TFLITE | float | Qualcomm® QCS9075 | 3.056 ms | 3 - 9 MB | NPU |
| SESR-M5 | TFLITE | float | Qualcomm® QCS8750 | 1.03 ms | 0 - 24 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.456 ms | 0 - 31 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.861 ms | 0 - 38 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® QCS6490 | 2.601 ms | 0 - 3 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® QCS8275 | 1.988 ms | 0 - 23 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.622 ms | 0 - 2 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® SA8775P | 21.759 ms | 4 - 26 MB | GPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® SA8650P | 21.759 ms | 4 - 26 MB | GPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® SA8255P | 21.759 ms | 4 - 26 MB | GPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® QCS8450 | 0.861 ms | 0 - 38 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.307 ms | 0 - 24 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.79 ms | 0 - 21 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® QCM6690 | 6.851 ms | 3 - 137 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® QCS9075 | 0.899 ms | 0 - 3 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® SA7255P | 1.988 ms | 0 - 23 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Snapdragon® 8 Elite Mobile | 0.38 ms | 0 - 24 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® SA8295P | 1.363 ms | 0 - 22 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® QCS7790 | 0.79 ms | 0 - 21 MB | NPU |
| SESR-M5 | TFLITE | w8a8 | Qualcomm® QCS8750 | 0.38 ms | 0 - 24 MB | NPU |
License
- The license for the original implementation of SESR-M5 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.
