Dreambooth is one of the method of finetune the pretrained text-to-image model.Given as input just a few images of a subject, it learns to bind a unique identifier with that specific subject.

AltDiffusion which is a multilingual text-to-image model.It currently supports 9 languages and will support more in future. We can use the dreambooth to finetune the AltDiffusion,so that we can get a multilingual Dreambooth model which could supports 9 languages.

Here we give a example model finetuned use a dozen pictures of Uzumaki Naruto downloaded from web.The example code of inference is shown bellow.You can have a try and maybe train your own dreambooth.Hopes have fun!

Example code of inference

from diffusers import AltDiffusionPipeline, DPMSolverMultistepScheduler
import torch

pipe = AltDiffusionPipeline.from_pretrained("BAAI/DreamBooth-AltDiffusion")
pipe = pipe.to("cuda")

pipe.scheduler = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config)

prompt = "一张<鸣人>男孩的照片"
image = pipe(prompt, num_inference_steps=25).images[0]
image.show()