Sign In

Rambi-XL SickLizard

11
127
190
5
Verified:
SafeTensor
Type
LoRA
Stats
113
190
Reviews
Published
Dec 12, 2023
Base Model
SDXL 1.0
Training
Steps: 3,360
Epochs: 16
Usage Tips
Clip Skip: 2
Strength: 1
Trigger Words
r4mb1
animated
hyperrealistic
pixel-art
16-bit
comic-style
cartoon
muppet
plushie
funko
+9 more
Training Images
Download
Hash
AutoV2
0930B1F727

Hunters killed his mother when he was a child.

But he´s not a child anymore!

And this time, he is the hunter!

You know his name
v2.0:
Massively improved dataset, more styles, more variety. Funko toys, Copics sketch,
Play around with weighting facial expressions and art styles.
Works on many models. Have fun. C&C appreciated

SDXL 1.0 Clip Skip 2 CFG 7 LoRA Strength: 0.7 - 1.1

Trigger: R4mb1

other tags:

animated, hyperrealistic, comic-style, angry,
crazy, knife, night, smiling, yelling,
gun, cartoon, laughing, jumping off,
helicopter, bonfire, muppet, plushie,
pixel-art, 16-bit, desert, forest,
burning jungle, copics, landscape,

funko,

----------------------------------------------------------------------------
v1.0:
I had that idea in my head for at least 15 years now.
Stable Diffusion failed to get me the right results but Dall E 3 actually did a great job.
Problem is Bing lets me only do so many generations per day so in this model, I only had 15 training images.

Dall E 3 original prompt:
>>"Rambi": Bambi, reminagined as an armed masculine muscular Rambo animated character. Wearing red headband, Jheri curls, grim, stoic expression. Full body shot.<<

So this model can definitely get improved with more and better input.

SDXL 1.0 Clip Skip 2 CFG 7 LoRA Strength: 0.7 - 1.1

Trigger: R4mb1

other tags:

comic style
hyperrealistic
armed
angry
full body

I tried not to overtrain the model but due to the given data, that was thin line between not getting any Rambi and getting basicly the training images back.

Feel free to send me any "Rambi" images that are capable of improving the dataset :--)