rexroth0619

Noticed in original Diffusers colab there's a section saying "training text encoder has a much better result." Just wondering where should I enable text encoder training with current implementation? Thanks

Train concept other than person?

2

Newcomer here, may sound like a dumb question -- 1. How do I train something else other than a person? I notice in regularization images section there is only images...

[Question] Should we just use "INSTANCE_NAME" in the prompt when generating new concept?

1

Might seem like a dumb question, but what prompt format should be used when we generate images for a newly trained concept? e.g. In the original dreambooth implementation we use...

Colab last cell shows AttributeError: module 'PIL.Image' has no attribute 'Resampling'

3

Hi, first time running this in colab, ran every cell as instructed in colab yet last cell can't process and shows AttributeError: module 'PIL.Image' has no attribute 'Resampling' Please advise