About this model
This is a LoRA extracted with permission from pmejna's fantastic graphic-art model. This was extracted against SD 2.1 768, at dim 300 and float16. Results versus the graphic-art checkpoint are similar but not identical when used with SD 2.1 768.I may extract a higher dim version to see if it works better than this version. Well, that gave me an error, so maybe not.
Copyrights, Licensing and support
If you like this LoRA and pmejna's original model and you want to support their current and upcoming projects: here is their coffee jar
As pmejna has mentioned on their other models - you can use it. Just keep in mind that they spent their money and time to make it. So if you like it, go give them your support! :)
The only restriction is selling their models.
Installation:
Simply put this LoRA model in the models/lora folder of your SD installation.
Versions:
There is one version of this model so far. No keyword - you can use it however you like. This is the alpha version.
Description
FAQ
Comments (7)
good
RuntimeError: mat1 and mat2 shapes cannot be multiplied (77x768 and 1024x300)
That's strange, as it works on my end. I'll have to take a look at it.
RuntimeError: mat1 and mat2 shapes cannot be multiplied (77x768 and 1024x300)
It's trained on 2.1
do not work for 1.5 / 2.1 models. RuntimeError: Given groups=1, weight of size [300, 320, 1, 1], expected input[1, 1, 8400, 320] to have 320 channels, but got 1 channels instead
That's strange, as it works on my end. I'll have to take a look at it.