NOTICE:
This LoRA was trained specifically for Pony Diffusion V6 XL.
I can NOT guarantee this works with any other SDXL models.
This LoRA was trained using 1024x1024 resolution images.
Some tags i Recommend with this LoRA are:
White Fur
Smug
White Ears
The actual trigger word for this LoRA is BoykisserFur
(For the V2 model, I highly suggest using 0.95 strength)
Another issue i run into sometimes is that the LoRA doesn't seem to work.
If this happens for you, put the strength at 0.9
This tends to fix the issue for me.
(Although this could just be an issue with my setup)
Description
FAQ
Comments (8)
why is this LoRA almost ~2GBs? that's pretty comparable in size to most pruned model checkpoints.
Is there any chance you could shrink and/or prune the model any further?
This is being worked on, im looking to see if i can make a smaller sized lora ^^
New versions have been added for reduced file size :3
Cool but the other model gets the face more accurate
That is correct, sadly the original is 1,7gb which is too much for most people. I made a resized version which is a bit smaller with minimal impact to quality / changes. V2 was retrained specifically for smaller file size for those wanting it. I prefer the original myself too :3
@YumeReborn Wait, are you talking about the 1.0 of this model or the SD 1.5 model? Because I am talking about the SD 1.5 model.
Rah
I'm carrying how much this gets used on my back despite popping in once or twice every month