Mario Larcher

I add also that the same job, once I tried to stop it, it remained pending for a lot of time, it took more than one day to stop and...

OK, I think I found the problem. ```python from datasets import load_dataset dataset = load_dataset("nelorth/oxford-flowers") image = dataset["train"][0]["image"] print(image.keys()) ``` gives you: `dict_keys(['bytes', 'path'])` This breaks `hf_datasets_augs_helper` that expect `dataset["train"][0]["image"]`...

Sure, it's not a big deal, maybe I'd consider adding something like `use_norm=True` or `norm=True` in the `MappingFeedForwardBlock`. It'll keep things as they are for now, but later on, if...

My understanding is that you use `aug_cond` when you wish to provide the model with information about the augmentations using Fourier Features: https://github.com/crowsonkb/k-diffusion/blob/6ab5146d4a5ef63901326489f31f1d8e7dd36b48/k_diffusion/models/image_transformer_v2.py#L657 https://github.com/crowsonkb/k-diffusion/blob/6ab5146d4a5ef63901326489f31f1d8e7dd36b48/k_diffusion/models/image_transformer_v2.py#L658 https://github.com/crowsonkb/k-diffusion/blob/6ab5146d4a5ef63901326489f31f1d8e7dd36b48/k_diffusion/models/image_transformer_v2.py#L718 On the other hand, if...

My problem for ch4 was that I had the two input files but they were empty, after replacing them with the correct files everything worked.

Hi @kohya-ss and all! I've created a script for resizing and cropping images for bucketing. Given the discussion here, I believe it can be useful as a quick pre-processing step....