README.md

---
pipeline_tag: image-to-image
---

# SuperResolution

This version of SuperResolution has been converted to run on the Axera NPU using **w8a8** quantization.

This model has been optimized with the following LoRA: 

Compatible with Pulsar2 version: 4.2

## Convert tools links:

For those who are interested in model conversion, you can try to export axmodel through 

- [The github repo of AXera Platform](https://github.com/AXERA-TECH/SuperResolution.axera)

- [How to Convert ONNX to axmodel](https://github.com/AXERA-TECH/SuperResolution.axera/tree/master/model_convert)

- [Pulsar2 tools](https://huggingface.co/AXERA-TECH/Pulsar2)


## Support Platform

- AX650
  - [M4N-Dock(爱芯派Pro)](https://wiki.sipeed.com/hardware/zh/maixIV/m4ndock/m4ndock.html)
  - [M.2 Accelerator card](https://axcl-docs.readthedocs.io/zh-cn/latest/doc_guide_hardware.html)
  
|Chips|model|cost|
|--|--|--|
|AX650|EDSR|800 ms|
|     |ESPCN|22 ms|

## How to use

Download all files from this repository to the device

```

root@ax650:~/SuperResolution# tree -L 1
.
├── assert
├── config.json
├── experiment
├── model_convert
├── python
├── README.md
└── video

6 directories, 2 files
```

### Requirements

```
pip install -r python/requirements.txt
```

pyaxengine 是 npu 的 python api，详细安装请参考: https://github.com/AXERA-TECH/pyaxengine


### Inference

Input Data:

```
└── video
    └── test_1920x1080.mp4
```

#### Inference with AX650 Host, such as M4N-Dock(爱芯派Pro)

```
root@ax650 ~/SuperResolution # python python/run_axmodel.py --model model_convert/axmodel/edsr_baseline_x2_1.axmodel --scale 2 --dir_demo video/test_1920x1080.mp4
[INFO] Available providers:  ['AxEngineExecutionProvider']
[INFO] Using provider: AxEngineExecutionProvider
[INFO] Chip type: ChipType.MC50
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Engine version: 2.12.0s
[INFO] Model type: 2 (triple core)
[INFO] Compiler version: 4.2 6bff2f67
100%|█████████████████████████████████████████| 267/267 [10:06<00:00,  2.27s/it]
Total time: 99.582 seconds for 267 frames
Average time: 0.373 seconds for each frame

```

The output file in `experiment/test_1920x1080_x2.avi`
![Example Image](video/2.png)

Output Data:

```
├── experiment
│   └── test_1920x1080_x2.avi
```

#### Inference with M.2 Accelerator card

```bash
$ cd python
$ python gradio_demo.py

[INFO] Available providers:  ['AXCLRTExecutionProvider']
[INFO] Using provider: AXCLRTExecutionProvider
[INFO] SOC Name: AX650N
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Compiler version: 4.2 6bff2f67
[INFO] Using provider: AXCLRTExecutionProvider
[INFO] SOC Name: AX650N
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Compiler version: 4.2 6bff2f67
==================================================
🌐 SuperResolution 超分辨率 Web UI 已启动！
🔗 本地访问: http://127.0.0.1:7860
🔗 局域网访问: http://10.126.33.124:7860
==================================================
```
Then open the link in the browser to use the web UI:

![gradio_demo](./assert/gradio_demo.jpg)