MakotoShinkai-Style-Diffusion

MakotoShinkai-Style-Diffusion

This is the fine-tuned Stable Diffusion model trained by dreambooth with 10k steps, images dataset from MakotoShinkai's videos. Use the tokens MakotoShinkaiStyle in your prompts for the effect.

🧨 Diffusers

This model can be used just like any other Stable Diffusion model. For more information, please have a look at the Stable Diffusion.

You can also export the model to ONNX, MPS and/or FLAX/JAX.

from diffusers import StableDiffusionPipeline
import torch
import os
from PIL import Image
model_id = "michaelz/MakotoShinkai-Style-Diffusion"
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float32)
pipe.requires_safety_checker = False
pipe.safety_checker = None
pipe = pipe.to("cuda")

save_dir = os.path.join("results")
os.makedirs(save_dir,exist_ok=True)
prompt_list = [
               "MakotoShinkaiStyle, a beautiful sky",
               "MakotoShinkaiStyle, a beautiful landscape",
               "MakotoShinkaiStyle, a beautiful city"]
for prompt in prompt_list:
    torch.manual_seed(-1)
    image = pipe(prompt,width = 512,height = 512).images[0]
    save_path = os.path.join(save_dir,prompt + ".png")
    image.save(save_path)

Sample images from the model:

Version 1 (MakotoShinkai-Style-Diffusion-10k): This model was trained by dreambooth by 10k steps