ERROR conda.core.link:_execute(507): An error occurred while installing package 'conda-forge::async_generator-1.10-py_0'

[Rafiur]

While creating the environment this error occurs and rolls back the creation of all other libraries. How do I get past this? Also if I wanted to train my own dataset what is the pipeline?

[CrohnEngineer]

Hey @Rafiur ,

thanks for passing by! Seems like you're trying to install as a package "conda-forge", which instead is a channel for downloading packages. To replicate our environment just clone our repo, and then from the bash execute the following lines:

$ conda env create -f environment.yml
$ conda activate icpr2020

On what kind of dataset do you want to train on? Please be aware that this is a repository for replicating the results of a scientific paper, so the pipeline is tailored to the two datasets we have used (i.e., Deep Fake Detection Challenge and FaceForensics++). Bests,

Edoardo

[Rafiur]

The error still occurs when I run "conda env create -f environment.yml" on ananconda prompt

yes I have seen that the repository is tailored for two datasets but I was hoping to know the preprocessing procedure of the dataset and how to do that

[nicobonne]

Hey @Rafiur, I just managed to recreate the environment from scratch under Ubuntu 20.10, so I think the error is related to Windows specific packages managed by Anaconda. I don't have a solution for you, try updating/reinstalling Anaconda hoping for an improvement. In general, I saw many people here using Windows without this problem, so my guess is that there is something broken in your Anaconda configuration. Maybe @CrohnEngineer has some Windows specific hints to share.

Regarding the training with your dataset: we don't cover this in our scripts. If you are really motivated in doing so, the steps to follow are: 1) indexing videos by modifying one of the index scripts index_dfdc.py or index_ffpp.py 2) extracting faces by modifing extract_faces.py

then it should be straightforward using one of the training scripts and test them accordingly.

[Rafiur]

Yes, it seems it was an issue on my end as I have done the same on another windows device and no error occurred. I will try to do so following the given steps. Thank you for the valuable feedbacks.