pink hair guy in glasses, photograph, sporty body, cinematic lighting, clear eyes, perfect face, blush, beautiful nose, beautiful eyes, detailed eyes
Anime Diffusion
Anime Diffusion2 is a latent text-to-image diffusion model based on Vintedois (22h) Diffusion model trained on BLIP captions for Danbou set + Demon slayer + arts from 4ch
Model Description
License
This model is open access and available to all, with a CreativeML OpenRAIL-M license further specifying rights and usage. The CreativeML OpenRAIL License specifies:
- You can't use the model to deliberately produce nor share illegal or harmful outputs or content
- The authors claims no rights on the outputs you generate, you are free to use them and are accountable for their use which must not go against the provisions set in the license
- You may re-distribute the weights and use the model commercially and/or as a service. If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the CreativeML OpenRAIL-M to all your users (please read the license entirely and carefully) Please read the full license here
Downstream Uses
This model can be used for entertainment purposes and as a generative art assistant.
Example Code
import torch
from torch import autocast
from diffusers import StableDiffusionPipeline
pipe = StableDiffusionPipeline.from_pretrained(
'AlexWortega/AnimeDiffuion2',
torch_dtype=torch.float32
).to('cuda')
negative_prompt = """low-res, duplicate, poorly drawn face, ugly, undetailed face"""
d = 'white girl'
prompt = f"pink hair guy in glasses, photograph, sporty body, cinematic lighting, clear eyes, perfect face, blush, beautiful nose, beautiful eyes, detailed eyes"
num_samples = 1
with torch.inference_mode():
images = pipe([prompt] * num_samples,
negative_prompt = [negative_prompt]*num_samples,
height=512, width=512,
num_inference_steps=50,
guidance_scale=8,
).images
images[0].save("test.png")
Team Member and Acknowledgements
This project would not have been possible without the incredible work by the CompVis Researchers.
In order to reach me, here is my blog: