Text-to-Image
Diffusers
PyTorch
StableDiffusionPipeline
stable-diffusion
diffusion-models-class
dreambooth-hackathon
food
import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("kokuma/mazapan", dtype=torch.bfloat16, device_map="cuda")
prompt = "a cute bunny, mazapan"
image = pipe(prompt).images[0]
DreamBooth model for the mazapan concept trained by kokuma on the kokuma/figuritas-de-mazapan dataset.
This is a Stable Diffusion model fine-tuned on the mazapan concept with DreamBooth for the food theme.
This model was created as part of the DreamBooth Hackathon 🔥. Visit the organisation page for instructions on how to take part!
Prompts
- a cute X, mazapan:
a cute bunny, mazapan - a cute X made of mazapan:
a cute robot made of mazapan - a photograph of a cute X, mazapan:
a photograph of a cute dog, mazapan
Suggested parameters
- CFG scale: Between 6 and 8
- Samplers: Euler a, Euler, DPM2 a, DPM++ SDE, DPM fast, DPM adaptive, DPM2 a Karras
Examples
| a cute dog, mazapan | a cute sparrow, mazapan | a cute bear, mazapan |
|---|---|---|
 |
 |
 |
| a cute koala, mazapan | a cute robot made of mazapan | a cute fox, mazapan |
|---|---|---|
 |
 |
 |
Usage
from diffusers import StableDiffusionPipeline
pipeline = StableDiffusionPipeline.from_pretrained('kokuma/mazapan')
image = pipeline().images[0]
image
- Downloads last month
- -