license: creativeml-openrail-m
tags:
- pytorch
- diffusers
- stable-diffusion
- text-to-image
- diffusion-models-class
- dreambooth-hackathon
- animal
widget:
- text: a photo of a pssg wearing goggles
- text: a photo of a pssg sugar glider wearing goggles
DreamBooth model for the pssg concept trained by pharmapsychotic
Sugar gliders are adorable creatures! I've never had one as a pet but I've been tempted. Imagine having one in your shirt pocket and feeding it snacks as you work. 😍
Anyway, I created a few AI renders of sugar gliders and mixed in with some photos of the critters and trained a model for the DreamBooth Hackathon! If you enjoy the model or just find the results funny and cute, drop a like on the model!
To use the model be sure to include pssg in your prompt (PharmapSychotic Sugar Glider) or pssg sugar glider for a stronger effect. I recommend using a version of the inference that has cross attention control so you can balance the influence of the sugar glider and the weird scenarios you put him in. I trained to 10,000 steps and it overcooked so dropped back to the 2,500 step checkpoint but still need to boost other things in the prompts like (((goggles))) to overcome the default pssg influence.
See below for usage!
Examples
 |
 |
 |
 |
 |
 |
 |
 |
 |
Usage
With Diffusers
from diffusers import StableDiffusionPipeline
pipeline = StableDiffusionPipeline.from_pretrained('pharma/sugar-glider')
image = pipeline().images[0]
image
With SD Web UI
To use with SD Web UI download sugar_gliders_pssg_2500.ckpt and put in your models folder.