If you'd like to support me and do more:
https://ko-fi.com/edgonkofi - commissions open!
If you're looking for a >>LoRA Making Tutorial<<
Fantasy Armors - by EDG
I honestly dislike reposting.
But since many of my fantasy armor models are overlooked I have to do this...
I hope some of you will discover new concepts this way!
Cheers!
PS: Make sure to have to LoCon extension installed for this one, some of these armors are LoCon.
Description
FAQ
Comments (19)
had this trigger;
stable-diffusion-webui\extensions-builtin\Lora\lora.py", line 192, in load_lora
assert False, f'Bad Lora layer name: {key_diffusers} - must end in lora_up.weight, lora_down.weight or alpha'
AssertionError: Bad Lora layer name: lora_unet_down_blocks_0_downsamplers_0_conv.lora_mid.weight - must end in lora_up.weight, lora_down.weight or alpha
Mmmmh
Many users seem to have that issue lately.
Before you could mix LoRA and LoCon just fine. But not anymore it seems...
Try putting the model into your lycoris folder if you have one and install the locon extension if you haven't yet.
@EDG yeah gave that a try;
weight's shape is different: lora_unet_input_blocks_3_0_op.lora_down.weight expected torch.Size([30, 320, 3, 3]) found torch.Size([30, 320]). SD version may be different
Testing now
@Roscosmos it's so frustrating when people update stuff and break it :/
Sorry you have to go through that
@EDG no worries!, such is life.
@EDG Got it working!, I think?!?, had to put in lycoris folder.
same here sir
@BigFatEvilCat yeah must be moved to the Lycoris folder for it to trigger.
@Roscosmos Thank you for the tip! Will do it after work. :)
GILDED ARMOR NEXT
Civitai is trying to be smart with it's webui extension and sends my images to the original model, without displaying them here for some reason.
weird!
Hey, i'm big fan of your loras thanks for releasing such amazing lora, I have one question about training lora, i found your lora training guide, and it was amazing, i want to train clothes also, i captioned everything as in your guide says, but i have no idea about training setting such as parameters and steps and epoch settings after captioning, I don't know anything about it, would be very helpful if you be able to give me a source or a video to follow, or maybe a premade lora training script, Thank you so much
Hello!
Thank you for enjoying my content, I am glad.
I will write an article about it and release it soon. Giving you the information you need.
If you are in hurry and can't wait, I recommend you head to https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/main/kohya-LoRA-dreambooth.ipynb
There you'll be able to train your LoRA on Google's cloud
All the parameters are pre-entered so it's easy!
Tale care and have fun!
@EDG I'll definitely wait, Thank you so much again
Thanks for you beautiful work. Where do I find LoCon? It is not in my Extension
a1111-sd-webui-lycoris
And for the next release: SAMURAI ARMOR hehe
Goth Variant us bugged
something wrong with layername
also iron corset
:
AssertionError: Bad Lora layer name: lora_unet_down_blocks_0_downsamplers_0_conv.lora_mid.weight - must end in lora_up.weight, lora_down.weight or alpha
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.