Ivan Skorokhodov

Hi! I am sorry for answering that late. Yes, we indeed preprocessed the dataset (as specified in the Experiments section) with the procedure described in Section 3.3/Appendix C/Algorithm 1. [Here](https://gist.github.com/universome/3140f74058a48aa56a556b0d9e24e857)...

Hmm, the currently released dataset has images sorted by their InceptionV3 likelihood (from least to most probable). I think in the above script, one should also shuffle the images. I...

Hi! Yes, we plan to release the dataset by the end of the next week

Hi! Our appendix is concatenated to the main part of the paper in its arxiv version. You can get it by scrolling below the references section: https://arxiv.org/abs/2104.06954

@Devetec yes, you can run the model in Colab: [here is the Colab version](https://colab.research.google.com/drive/1XMZo1K3Qz24_n4o_xY4SSBHeo6ooS_Ak?usp=sharing) of `notebooks/generate.ipynb`.

@gllmp Could you please tell what problem you encounter when using your own .pkl file?

@gllmp Yes, as @Devetec said, it is not possible to use original StyleGAN2's checkpoints with ALIS code since there are large differences in architectures (like patchwise generation and coordinate conditioning)....

Hi! In my head, it was supposed to look the following way. You sample 9 anchors and assemble them into a 3x3 grid (right now there are 3 anchors sampled...

I agree that the code is not too transparent and might take a lot of time to get one's head around. I checked several places that I think are the...

The problem arises because `grid` variable is not a normal `torch.Tensor`, but rather a tweaked `Distribution` variable. To fix the issue, I had to tweak lines: in `train_fns.py`: from ```...