Giters

LoSealL / VideoSuperResolution

A collection of state-of-the-art video or single-image super-resolution architectures, reimplemented in tensorflow.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Refactor of this project planned

LoSealL opened this issue · comments

commented

Refactor is planned. The objects are:

  • Support TF2.0
  • Unify calling of TF model and PyTorch model
  • Easier PIPY package to call from another project
  • Open access to pre-trained models
  • New models integrated from CVPR2019 and ICCV2019
  • More documents

Any kind of help is highly appreciated and welcome :)

commented

Hi, I have utilized your project within a colab notebook. it is working very nicely, thank you very much for sharing your project. I would love to help. maybe I can pre-train some models or try to implement a new model? I would really like to try this one: TDAN https://arxiv.org/pdf/1812.02898.pdf but I don't know where to start...

commented

Hi @mesw, I very much appreciate your help!
First of all, you can start within VSRTorch, by simply writing tdan.py into VSRTorch/Models/;
Then link your model by adding an entry into VSRTorch/Models/__init__.py;

Those README.md files could do a little help as well.

commented

Your work is amazing! Thank you so much. RLSP is also a better model than FRVSR in VSR. Could you consider include it in your framework? https://arxiv.org/abs/1909.08080

commented

@Easoncyx Thanks, I will take a look

commented

I am so much thrilled to see refactor of this repo !
Can't wait.
So much thanks for presenting the VSR repo.