---
license: creativeml-openrail-m
language: 
  - en
thumbnail: "https://huggingface.co/Norod78/sd-simpsons-model/raw/main/examples/00496-2202810362-A%20beautiful%20hungry%20demon%20girl,%20John%20Philip%20Falter,%20Very%20detailed%20painting,%20Mark%20Ryden.jpg"
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
datasets:
- Norod78/simpsons-blip-captions
inference: true
---

# Simpsons diffusion
*Stable Diffusion fine tuned on images related to "The Simpsons"

If you want more details on how to generate your own blip cpationed dataset see this [colab](https://colab.research.google.com/gist/Norod/ee6ee3c4bf11c2d2be531d728ec30824/buildimagedatasetwithblipcaptionsanduploadtohf.ipynb)

Training was done using a slightly modified version of Hugging-Face's text to image training [example script](https://github.com/huggingface/diffusers/blob/main/examples/text_to_image/train_text_to_image.py)

## About

Put in a text prompt and generate cartoony/simpsony images

**A beautiful hungry demon girl, John Philip Falter, Very detailed painting, Mark Ryden**

![A beautiful hungry demon girl, John Philip Falter, Very detailed painting, Mark Ryden](https://huggingface.co/Norod78/sd-simpsons-model/raw/main/examples/00496-2202810362-A%20beautiful%20hungry%20demon%20girl,%20John%20Philip%20Falter,%20Very%20detailed%20painting,%20Mark%20Ryden.jpg)

**Gal Gadot, cartoon**

![Gal Gadot, cartoon](https://huggingface.co/Norod78/sd-simpsons-model/raw/main/examples/00323-2574793241-Gal%20Gadot,%20cartoon.jpg)

## More examples

The [examples](https://huggingface.co/Norod78/sd-simpsons-model/tree/main/examples) folder contains a few images generated by this model's ckpt file using [stable-diffusion-webui](https://github.com/AUTOMATIC1111/stable-diffusion-webui) which means their EXIF info contain the parameter used to generate them

## Sample code


```py
from diffusers import StableDiffusionPipeline, LMSDiscreteScheduler
import torch

# this will substitute the default PNDM scheduler for K-LMS  
lms = LMSDiscreteScheduler(
    beta_start=0.00085, 
    beta_end=0.012, 
    beta_schedule="scaled_linear"
)

guidance_scale=9
seed=7777
steps=100

model_id = "Norod78/sd-simpsons-model"
pipe = StableDiffusionPipeline.from_pretrained(model_id, scheduler=lms, torch_dtype=torch.float16)
pipe.to("cuda")

prompt = "a photo of an astronaut riding a horse on mars"
image = pipe(prompt).images[0]  
    
image.save("astronaut_rides_horse.png")
```

## Dataset and Training

Finetuned for 10,000 iterations upon [Runway ML's Stable-Diffusion v1.5](https://huggingface.co/runwayml/stable-diffusion-v1-5) on [BLIP captioned Simpsons images](https://huggingface.co/datasets/Norod78/simpsons-blip-captions) using 1xA5000 GPU on my home desktop computer

Trained by [@Norod78](https://twitter.com/Norod78)