d8ahazard

I'd start from scratch, re-install xformers following the official Auto1111 instructions, and ensure you're on the latest version of the extension.

> I'd love this, but would rather wait for it to land in webui's traning system first, and hopefully we can reuse the preprocessing part. I don't think it necessarily...

> There are some training scripts: https://github.com/bmaltais/kohya_ss/blob/master/diffusers_fine_tuning/fine_tune.py https://github.com/harubaru/waifu-diffusion/blob/main/trainer/diffusers_trainer.py > > Don't know how much we can learn from them. Honestly, in looking at this, it seems like it's basically just...

So, an update to this: I haven't implemented bucketing...yet...but I've started working on a new Dataset class that should allow us to deal with multiple "sets" of images more intelligently....

Implemented!

Wizards are not launched, per-se, it's just running some calculations. Buttons have been moved, and better output should be shown in the UI to indicate "things happened".

I'll need to devise a way to detect whether or not the model is 512 or 768 and copy the proper config file. Ideally, the webui would just do this...

Because we shouldn't need it. I'm actively working on comparing the 512 and 768 checkpoints to see which keys we can use to figure this out. I'm already detecting whether...

I'm having some difficulty reproducing the issue. I took the -v config file and loaded the 512 model with it, and had no weird outputs. Could you possibly provide a...

Iz good, checkbox fixed. Da!