DominikDoom

@Cyberes > > manually resume training from a previous checkpoint > > Are you doing this with `train_dreambooth.py`? Yes, the script takes `--pretrained_model_name_or_path` as an argument. In most Colabs etc....

Thanks for the suggestion, but I don't think this is the better way. While it resolves some issues like spaces cancelling the edit, it also creates many more that are...

I'm not able to reproduce this locally and the UI looks a bit different as well, could you tell me which colab you are using?

I just tried it with that colab, but it looks normal for me (and also like the current local version, which your screenshot doesn't match):  Are you maybe using...

> Besides the tagcomplete option shows rather slow than before, but I suppose it's some performance issue of webui itself, or the extension system (I'm not sure but the slow...

Some models definitely benefit more from it, but since a lot of tags are pretty general, they can also be used in the standard 1.4 model with decent results (for...

Well, it definitely wouldn't hurt to have the option, I'm just not sure how I want to include it in the config. The popular custom models are nearly all trained...

Closing this since it's resolved now. (See #206 which was implemented in https://github.com/TheLastBen/diffusers/commit/0b77c4e56bf377d424c7ed87d6cb1804469b0217).

Swallowing characters has never happened to me before, even when resultStepLength is a much higher value like 50000. Yes, the browser will lag for potentially seconds in that case -...

Alright, thanks for the details. I'll try it on Firefox and see if I can find the cause.