README.md

---
library_name: pytorch
license: other
tags:
- android
pipeline_tag: image-to-image

---

![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/web-assets/model_demo.png)

# QuickSRNetMedium: Optimized for Qualcomm Devices

QuickSRNet Medium is designed for upscaling images on mobile platforms to sharpen in real-time.

This is based on the implementation of QuickSRNetMedium found [here](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/quicksrnet).
This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/quicksrnetmedium) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).

Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) 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 2.37, ONNX Runtime 1.23.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.46.0/quicksrnetmedium-onnx-float.zip)
| QNN_DLC | float | Universal | QAIRT 2.42 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.46.0/quicksrnetmedium-qnn_dlc-float.zip)
| QNN_DLC | w8a8 | Universal | QAIRT 2.42 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.46.0/quicksrnetmedium-qnn_dlc-w8a8.zip)
| TFLITE | float | Universal | QAIRT 2.42, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.46.0/quicksrnetmedium-tflite-float.zip)
| TFLITE | w8a8 | Universal | QAIRT 2.42, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.46.0/quicksrnetmedium-tflite-w8a8.zip)

For more device-specific assets and performance metrics, visit **[QuickSRNetMedium on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/quicksrnetmedium)**.


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/quicksrnetmedium) 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 [QuickSRNetMedium on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/quicksrnetmedium) for usage instructions.

## Model Details

**Model Type:** Model_use_case.super_resolution

**Model Stats:**
- Model checkpoint: quicksrnet_medium_3x_checkpoint
- Input resolution: 128x128
- Number of parameters: 61.0K
- Model size (float): 243 KB
- Model size (w8a8): 73.9 KB

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| QuickSRNetMedium | ONNX | float | Snapdragon® X Elite | 1.096 ms | 8 - 8 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.861 ms | 0 - 94 MB | NPU
| QuickSRNetMedium | ONNX | float | Qualcomm® QCS8550 (Proxy) | 1.178 ms | 0 - 37 MB | NPU
| QuickSRNetMedium | ONNX | float | Qualcomm® QCS9075 | 1.754 ms | 7 - 10 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.655 ms | 0 - 91 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.562 ms | 0 - 90 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® X Elite | 0.893 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 0.489 ms | 0 - 30 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 2.535 ms | 0 - 23 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 0.779 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA8775P | 1.172 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® QCS9075 | 1.191 ms | 0 - 5 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 1.151 ms | 0 - 31 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA7255P | 2.535 ms | 0 - 23 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA8295P | 1.587 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.381 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.372 ms | 0 - 24 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® X Elite | 0.445 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.231 ms | 0 - 27 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS6490 | 1.38 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS8275 (Proxy) | 0.817 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.339 ms | 0 - 1 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA8775P | 0.538 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 0.493 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCM6690 | 1.581 ms | 0 - 17 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS8450 (Proxy) | 0.517 ms | 0 - 27 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA7255P | 0.817 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA8295P | 0.722 ms | 0 - 16 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.177 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.375 ms | 0 - 16 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.148 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 0.724 ms | 0 - 30 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 3.09 ms | 3 - 25 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 1.097 ms | 0 - 1 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA8775P | 1.589 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® QCS9075 | 1.408 ms | 1 - 6 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 1.46 ms | 0 - 30 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA7255P | 3.09 ms | 3 - 25 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA8295P | 1.818 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.506 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.435 ms | 0 - 24 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.298 ms | 0 - 28 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS6490 | 1.139 ms | 0 - 3 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS8275 (Proxy) | 1.019 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.442 ms | 0 - 1 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA8775P | 0.638 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS9075 | 0.569 ms | 0 - 3 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCM6690 | 1.927 ms | 0 - 17 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS8450 (Proxy) | 0.512 ms | 0 - 29 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA7255P | 1.019 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA8295P | 0.838 ms | 0 - 16 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.223 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.465 ms | 0 - 17 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.176 ms | 0 - 21 MB | NPU

## License
* The license for the original implementation of QuickSRNetMedium can be found
  [here](https://github.com/quic/aimet-model-zoo/blob/develop/LICENSE.pdf).

## References
* [QuickSRNet: Plain Single-Image Super-Resolution Architecture for Faster Inference on Mobile Platforms](https://arxiv.org/abs/2303.04336)
* [Source Model Implementation](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/quicksrnet)

## Community
* Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
* For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).