!!! PLEASE READ INSTRUCTIONS !!!
This Lora aims to reproduce the claymation style from the popular show Celebrity Deathmatch aired on MTV back in the good old days
CLASSIC
the classic Celebrity Deathmatch look
usage:
Weight starts from 0.7
VERSATILE
this version is maybe less faithful to the original style but it has more variety in head shapes and this helps when used in conjunction with celeb LoRas to retain likeness
usage:
Weight starts from 0.8 up to 1.5
If you like what I do donate some Buzz or buy me a coffee here
If you want to post an image made with this LoRa on civitAi, please, do not cross-post but create a new post on this page with the ( + add Post) button, it will help me gain buzz without the need to rely on tipping.
Description
FAQ
Comments (24)
Bro! Seriously, this looks phenomenal!
Hey, thank you so much!
Oh, i'm about to go to town with this one!
I knew you would have liked this!
wow great job!
hey! thanks. I'll try to use some of your celebs LoRa to see how they interact with this one.
No way we got a model for PEAK
How do you use this? im using flux_dev.safetensors [2eda627c8a]
with the Lora provided but generated content is nowhere near the intended style.
any details how to get this going properly? im using automatic1111
I don't use a1111 so I can't help you, I'm sorry!
@Clumsy_Trainer Are you using comfyUI? Just wondering how I can get this to generate the style you set it to.
@OneDumbDude Yes, I use comfyUi. I just use rgthree LoRa loader (but you can use any other LoRa loader model) and play a bit with the weight, especially for VERSATILE
I almost got whiplash seeing this. Great job. No things has snapped me back to my childhood if the 90s quicker other than double dare. Thank you so much. Great work.
Glad you like it!
Now, here's something I haven't heard about in years. They definitely need to bring this back using the technology we have today.
Individually both are useful but I'm having trouble putting two character LoRAs into the same scene without it trying to blend them. Is there something I need to be doing in the prompt to prevent that?
And ironically one of the characters looks more classic with Versatile selected and vice versa.
sadly that's a problem with any diffusion model, always has been. You can test this by generating an image without the CDM LoRa, it'll blend the character.
For the character that looks different with versatile or classic, it depends on the celeb LoRa you are using and how they have been trained (specifically, how they are captioned and how strong the learning rate was during training).
You can achieve this by clever use of the img2img masking in a ComfyUI workflow or better yet use one of the custom nodes (also free like ComfyUI) that are designed for partitioning latent space so you can plug multiple different text encode prompts into different locations within the same latent space to exactly position where you want your multiple characters. This is one of the custom nodes that can be used for this (via ComfyUI) https://github.com/Davemane42/ComfyUI_Dave_CustomNode?tab=readme-ov-file
@AINickAskewAI Yes, there also are a couple of nodes that use the LoRa in a specific position and at a specific moment during steps.
Just Awsome! Thanks for creating!
You're welcome!
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.