import spaces
import os
import io
import IPython.display
from PIL import Image
import base64
import io
from PIL import Image
import gradio as gr
import requests
import time
import random
import numpy as np
import torch
import os
from transformers import ViTModel, ViTImageProcessor
from utils import text_encoder_forward
from diffusers import StableDiffusionPipeline, DPMSolverMultistepScheduler
from utils import latents_to_images, downsampling, merge_and_save_images
from omegaconf import OmegaConf
from accelerate.utils import set_seed
from tqdm import tqdm
from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion import StableDiffusionPipelineOutput
from PIL import Image
from models.celeb_embeddings import embedding_forward
import models.embedding_manager
import importlib
import time
import os
# os.environ['GRADIO_TEMP_DIR'] = 'qinghewang/tmp'
title = r"""
CharacterFactory: Sampling Consistent Characters with GANs for Diffusion Models
"""
description = r"""
Official Gradio demo for CharacterFactory: Sampling Consistent Characters with GANs for Diffusion Models.
How to use:
1. Enter prompts with the trigger words `a person`, and each line will generate an image.
2. You can choose to `create a new character` or `continue to use the current one`. We have provided some examples, click on the examples below to use.
3. You can choose to use the Normal version (the gender is random), the Man version, and the Woman version.
4. Click the Generate button to begin (Images are generated one by one).
5. Our method can be applied to illustrating books and stories, creating brand ambassadors, developing presentations, art design, identity-consistent data construction and more. Looking forward to your explorations!๐
6. If CharacterFactory is helpful, please help to โญ the Github Repo. Thanks!
"""
article = r"""
---
๐ **Citation**
If our work is helpful for your research or applications, please cite us via:
```bibtex
@article{wang2024characterfactory,
title={CharacterFactory: Sampling Consistent Characters with GANs for Diffusion Models},
author={Wang, Qinghe and Li, Baolu and Li, Xiaomin and Cao, Bing and Ma, Liqian and Lu, Huchuan and Jia, Xu},
journal={arXiv preprint arXiv:2404.15677},
year={2024}
}
```
๐ง **Contact**
If you have any questions, please feel free to open an issue or directly reach us out at qinghewang@mail.dlut.edu.cn.
"""
css = '''
#color-bg{display:flex;justify-content: center;align-items: center;}
.color-bg-item{width: 100%; height: 32px}
#main_button{width:100%}