search

sihyun-yu / digan

Official PyTorch implementation of Generating Videos with Dynamics-aware Implicit Generative Adversarial Networks (ICLR 2022).
https://sihyun.me/digan/
182 stars 19 forks source link

README file #13

Closed nicolgo closed 2 years ago

nicolgo commented 2 years ago

Following the README file, I failed to run the project. Here are some suggestions and questions:

  1. I think the author should tell us the project only support Linux system at least.
  2. I just want to train the model, but the guide is too simple. Firstly, what are the means of ""? Can I think it is just a temp name, it is ok to pass any text? Secondly, how to change the training options? Can I run the project successfully without any changes? Thirdly, where to place the data directory? data/UCF-101?
  3. it seems launch.py is just an encapsulation of train.py, since some default settings may not available for everyone, why not provide a set of train.py ?
  4. I have looked through the project, code quality is good, but the README is really a disaster.
sihyun-yu commented 2 years ago

Hi,

  1. Thank you for your suggestion. We did not add it since lots of deep learning implementation repositories assume the default system is built on Linux.
  2. EXP_NAME specifies the temporal directory to save checkpoints and several logs. We omitted the detailed explanation since such a term is similarly (and widely) used in similar project repositories, e.g., INR-GAN. Anyway, I feel sorry if you are confused about this. Regarding the data directory, you can manage it by modifying configs/main.yml. While we do not explicitly mention the data directory, but we mentioned, "You may change training options via modifying configs/main.yml" in the current README file. Considering the repositories is build a year ago and many people have already succeeded in reproducing the result, I think this is not a serious problem.
  3. This project is built upon the popular INR-GAN repository that uses launch.pyand train.py. It copies all of the executed code to the experiment directory.
nicolgo commented 2 years ago

Hi sihyun, Thanks for your time and explanation.