nitrosocke

/

spider-verse-diffusion

like 343

Yes, you can. Download the ckpt file : https://huggingface.co/nitrosocke/spider-verse-diffusion/resolve/main/spiderverse-v1-pruned.ckpt

And put it in the "models/Stable-diffusion" folder inside your AUTOMATIC1111 install. You'll be able to change model from the dropdown that is at the top left of your UI then

Same with any model mostly

hi, I tried that but since I'm using the spiderverse style model only I can't use mine. also tried merging them, but even then it doesn't detect well the spiderverse model

You cannot use 2 models at once, no. You can either try to merge (like you saw though, quality can vary) or you can train the spiderverse model on your face (lots more costly to do but best results). Lastly, you can just do img2img on your face and the spiderverse model, but it's not the same results for sure

i'm going to do the img2img, but how could i train the spiderverse model on my face? do i have to use the "training" option or is it other way? (sorry if these questions are dumb, but i'm new to these ai generated images)

This isn't the best place for such things mostly, not very interactive, you'd have more help on the discord. There are numerous guides out there on how to train on your face, it's the most popular subject relating to SD, look up "dreambooth" on youtube to get tutorials. The "train" tab isn't really good for this, you'll need to use either Dreambooth locally or in collab. "TheLastBen" or "Shivam" are the most popular and accessible right now, from what I get on the discord.

But once again, this has very few to do with the Spiderverse model now, and may be more appropriate on other internet forums and discords.

ok, thanks for the help!