app.py

#!/usr/bin/env python
# -*- coding:utf-8 -*-
# Power by Zongsheng Yue 2022-12-16 16:17:14

import os
import torch
import argparse
import numpy as np
import gradio as gr
from pathlib import Path
from einops import rearrange
from omegaconf import OmegaConf
from skimage import img_as_ubyte

from utils import util_opts
from utils import util_image
from utils import util_common

from sampler import DifIRSampler
from ResizeRight.resize_right import resize
from basicsr.utils.download_util import load_file_from_url

# setting configurations
cfg_path = 'configs/sample/iddpm_ffhq512_swinir.yaml'
configs = OmegaConf.load(cfg_path)
configs.aligned = False
configs.diffusion.timestep_respacing = '250'

# build the sampler for diffusion
sampler_dist = DifIRSampler(configs)

def predict(im_path, background_enhance, face_upsample, upscale, started_timesteps):
    assert isinstance(im_path, str)
    print(f'Processing image: {im_path}...')

    configs.background_enhance = background_enhance
    configs.face_upsample = face_upsample
    started_timesteps = int(started_timesteps)
    assert started_timesteps < int(configs.diffusion.params.timestep_respacing)

    # prepare the checkpoint
    if not Path(configs.model.ckpt_path).exists():
        load_file_from_url(
            url="https://github.com/zsyOAOA/DifFace/releases/download/V1.0/iddpm_ffhq512_ema500000.pth",
            model_dir=str(Path(configs.model.ckpt_path).parent),
            progress=True,
            file_name=Path(configs.model.ckpt_path).name,
            )
    if not Path(configs.model_ir.ckpt_path).exists():
        load_file_from_url(
            url="https://github.com/zsyOAOA/DifFace/releases/download/V1.0/General_Face_ffhq512.pth",
            model_dir=str(Path(configs.model_ir.ckpt_path).parent),
            progress=True,
            file_name=Path(configs.model_ir.ckpt_path).name,
            )

    # Load image
    im_lq = util_image.imread(im_path, chn='bgr', dtype='uint8')
    if upscale > 4:
        upscale = 4  # avoid momory exceeded due to too large upscale
    if upscale > 2 and min(im_lq.shape[:2])>1280:
        upscale = 2  # avoid momory exceeded due to too large img resolution
    configs.detection.upscale = int(upscale)

    if background_enhance:
        image_restored, face_restored, face_cropped = sampler_dist.sample_func_bfr_unaligned(
                y0=im_lq,
                start_timesteps=started_timesteps,
                need_restoration=True,
                draw_box=False,
                )   # h x w x c, numpy array, [0, 255], uint8, BGR
        image_restored = util_image.bgr2rgb(image_restored)
    else:
        image_restored = sampler_dist.sample_func_ir_aligned(
                y0=im_lq,
                start_timesteps=started_timesteps,
                need_restoration=True,
                )[0]  # b x c x h x w, [0, 1], torch tensor, RGB
        image_restored = util_image.tensor2img(
                image_restored.cpu(),
                rgb2bgr=False,
                out_type=np.uint8,
                min_max=(0, 1),
                )     # h x w x c, [0, 255], uint8, RGB, numpy array

    restored_image_dir = Path('restored_output')
    if not restored_image_dir.exists():
        restored_image_dir.mkdir()
    # save the whole image
    save_path = restored_image_dir / Path(im_path).name
    util_image.imwrite(image_restored, save_path, chn='rgb', dtype_in='uint8')

    return image_restored, str(save_path)

title = "DifFace: Blind Face Restoration with Diffused Error Contraction"
description = r"""
<b>Official Gradio demo</b> for <a href='https://github.com/zsyOAOA/DifFace' target='_blank'><b>DifFace: Blind Face Restoration with Diffused Error Contraction</b></a>.<br>
🔥 DifFace is a robust face restoration algorithm for old or corrupted photos.<br>
"""
article = r"""
If DifFace is helpful for your work, please help to ⭐ the <a href='https://github.com/zsyOAOA/DifFace' target='_blank'>Github Repo</a>. Thanks!
[![GitHub Stars](https://img.shields.io/github/stars/zsyOAOA/DifFace?affiliations=OWNER&color=green&style=social)](https://github.com/zsyOAOA/DifFace)

---

📝 **Citation**

If our work is useful for your research, please consider citing:
```bibtex
@article{yue2022difface,
  title={DifFace: Blind Face Restoration with Diffused Error Contraction},
  author={Yue, Zongsheng and Loy, Chen Change},
  journal={arXiv preprint arXiv:2212.06512},
  year={2022}
}
```

📋 **License**

This project is licensed under <a rel="license" href="https://github.com/zsyOAOA/DifFace/blob/master/LICENSE">S-Lab License 1.0</a>.
Redistribution and use for non-commercial purposes should follow this license.

📧 **Contact**
If you have any questions, please feel free to contact me via <b>zsyzam@gmail.com</b>.
![visitors](https://visitor-badge.laobi.icu/badge?page_id=zsyOAOA/DifFace)
"""

demo = gr.Interface(
    predict,
    inputs=[
        gr.Image(type="filepath", label="Input"),
        gr.Checkbox(value=True, label="Background_Enhance"),
        gr.Checkbox(value=True, label="Face_Upsample"),
        gr.Number(value=2, label="Rescaling_Factor (up to 4)"),
        gr.Slider(1, 160, value=100, step=10, label='Realism-Fidelity Trade-off')
    ],
    outputs=[
        gr.Image(type="numpy", label="Output"),
        gr.outputs.File(label="Download the output")
    ],
    title=title,
    description=description,
    article=article,
    examples=[
        ['./testdata/whole_imgs/00.jpg', True, True, 2, 100],
        ['./testdata/whole_imgs/01.jpg', True, True, 2, 100],
        ['./testdata/whole_imgs/04.jpg', True, True, 2, 100],
        ['./testdata/whole_imgs/05.jpg', True, True, 2, 100],
      ]
    )

demo.queue(concurrency_count=4)
demo.launch()