目录
一、拉取Deoldify项目代码
git地址:https://github.com/jantic/DeOldify
二、下载预训练模型
2.1、步骤
1)在项目根目录创建models文件夹
2)下载预训练模型文件,将三个文件上传至models目录
三、图片上色
from deoldify import device
from deoldify.device_id import DeviceId
from deoldify.visualize import *
from pathlib import Path
import os
os.environ['KMP_DUPLICATE_LIB_OK'] = "TRUE"
import warnings
warnings.filterwarnings("ignore", category=UserWarning, message=".*?Your .*? set is empty.*?")
#choices: CPU, GPU0...GPU7
device.set(device=DeviceId.GPU0)
plt.style.use('dark_background')
torch.backends.cudnn.benchmark=True
render_factor=35
result_path = None
p = Path('D:\\Gitlab\\DeOldify\\test_images')
all_file_path = []
count = 0
for file_name in p.rglob('*.jpg'):
all_file_path.append(file_name)
count+=1
for file_name in p.rglob('*.png'):
all_file_path.append(file_name)
count+=1
for file_name in p.rglob('*.jppeg'):
all_file_path.append(file_name)
count+=1
print(f"遍历待上色文件夹结束,共有{count}个图片文件\n")
# 遍历图片,上色
for img_file_path in all_file_path:
# arctan
colorizer = get_image_colorizer(artistic=True)
img_file_name = str(img_file_path).split('\\')[-1]
print(f"开始处理{img_file_name}")
colorizer.plot_transformed_image(path=img_file_path, render_factor=render_factor, compare=True)
print(f"{img_file_name}处理结束")
print("\n")
# 结束
print("图片上色程序结束")
四、视频上色
from deoldify import device
from deoldify.device_id import DeviceId
from deoldify.visualize import *
plt.style.use('dark_background')
import warnings
warnings.filterwarnings("ignore", category=UserWarning, message=".*?Your .*? set is empty.*?")
#choices: CPU, GPU0...GPU7
device.set(device=DeviceId.GPU0)
colorizer = get_video_colorizer()
#NOTE: Max is 44 with 11GB video cards. 21 is a good default
render_factor=21
result_path = None
file_name_ext = 'D:\\Gitlab\\DeOldify\\test_video\\test1.mp4'
colorizer.colorize_from_file_name(file_name_ext, render_factor=render_factor)