catboxanon
catboxanon
I think Divide & Bind would be the way to go about this now. https://sites.google.com/view/divide-and-bind
I'm going to close this because there is already several tools to do this, including one from a repo Auto personally recommends. https://github.com/kohya-ss/sd-scripts/blob/main/networks/extract_lora_from_models.py The training tab and related features to...
Closing as the script provided here does this.
Closing as stale.
Closing as stale.
Closing as duplicate of https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/6307 (I'll leave a follow-up comment there in a moment). Also see https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/6725 for a bit more background.
medvram or xformers are going to be incredibly more beneficial to both memory usage and speed than turning off batch cond/uncond.
Save this to the root directory of your install, run with path and type args specified. This will pre-calculate all the hashes. https://gist.github.com/catboxanon/4f56fdd1dd0ab207c6231eef3006c012
Would be resolved by https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/12350 This is also technically not a bug so I've reworded this and swapped the issue label.
There's been a lot of rework to how the webui initializes over the last few months and I think this is resolved now. If not, feel free to re-open or...