import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", dtype=torch.bfloat16, device_map="cuda")
pipe.load_lora_weights("Pclanglais/TintinIA")
prompt = "drawing of tintin riding giant snowydog, flying in the air, night sky with stars, close-up"
image = pipe(prompt).images[0]TintinIA is fine-tuned version of Stable-Diffusion-xl trained on 125 comics panels from Tintin album.
Currently TintinIA makes it possible to generate images of three characters: Tintin, Snowy and (to a lesser extent) Haddock. Best results are frequently obtained using "close-up". For some hard to draw characters, you can use our standard prefix "a comic panel of", although it usually comes at a in quality.
TintinIA is only released under a non-commercial license. It should be preferably used to create memes and parody content.
- Downloads last month
- 119
Model tree for Pclanglais/TintinIA
Base model
stabilityai/stable-diffusion-xl-base-1.0