Add EDSR and MDSR

[IMvision12]

Model description

EDSR (Enhanced Deep Residual Networks for Single Image Super-Resolution) is for image super resolution, here's the paper.

Open source status

• [x] The model implementation is available
• [x] The model weights are available

Provide useful links for the implementation

Official Implementation https://github.com/sanghyun-son/EDSR-PyTorch and https://github.com/LimBee/NTIRE2017

Your contribution

I'd like to work on incorporating this architecture into the HuggingFace. Please let me know if you think it's worth adding to huggingface.

@NielsRogge can you review this issue so that I can get started

[IMvision12]

@venkat-natchi

[venkat-natchi]

I'm going through the paper and the existing implementation. I will open the PR in few days.

[IMvision12]

I'm going through the paper and the existing implementation. I will open the PR in few days.

That's great, I have already gone through the paper and the architecture and started the implementation!

[venkat-natchi]

Sorry, I was away due to festival season here. I am done with the paper. Shall I start transforming this one into the HuggingFace model standards?

[IMvision12]

Go ahead and let me know if you need any help; I will be adding the MDSR model to the hub, and I believe the code should be similar to this Swin2sr PR also we can use it for reference.

[venkat-natchi]

Sure, thanks

[venkat-natchi]

Can we add both EDSR and MDSR in this PR?

[IMvision12]

No, I guess a separate PR is required.

[asrimanth]

Hello, can I work on this issue? Although I'm new to open-source contributions, I've worked on super-resolution models in the past and I was wondering why HuggingFace did not have these. I am familiar with PyTorch.

[venkat-natchi]

Thanks @asrimanth for the interest. I have an active PR going on for this issue.

19952

Kindly leave your comments there if you could.