【实战】轻轻松松使用StyleGAN（一）：创建令人惊讶的黄种人脸和专属于自己的老婆动漫头像

NVIDIA（英伟达）开源了StyleGAN，用它可以生成令人惊讶的逼真人脸；也可以像某些人所说的，生成专属于自己的老婆动漫头像。这些生成的人脸或者动漫头像都是此前这个世界上从来没有过的，完全是被“伟大的你”所创造出来的。

StyleGAN的源码可以到这里下载：

https://github.com/NVlabs/stylegan

有好事者把StyleGAN的论文翻译成中文，虽然阅读起来也并不是很容易懂，但还是比看英文论文要容易多了：

https://blog.csdn.net/a312863063/article/details/88761977

那么问题来了，怎样试一试StyleGAN的神奇魔力呢？

步骤如下：

（1）拥有或安装NVIDIA（英伟达）的显卡，比如：我的HP笔记本的操作系统是Windows 10，带有NVIDIA GeForce GTX 1060显卡，虽然配置并不高，也一样可以把StyleGAN跑起来；

（2）安装cuda 、cudnn、tensorflow-gpu 、 pillow等模块，建议使用Anaconda来安装，具体过程可以参考：

https://blog.csdn.net/weixin_41943311/article/details/91866987

https://blog.csdn.net/weixin_41943311/article/details/93747924

（3）下载StyleGAN源码：

下载完成后，把ZIP包解压缩，放到自己的工作目录下。

（4）下载模型文件：

有好事者已经把文件放到了百度网盘上，这样下载起来比较方便，

（4.1）已训练好的人脸模型（各个人种都有，但黄种人脸偏少，1024x1024）：karras2019stylegan-ffhq-1024x1024.pkl

百度网盘：https://pan.baidu.com/s/1ujItgpnHSw14Fw8I3Ai7Jw

提取码：ossw

（4.2）已训练好的动漫头像模型（日本动漫少女为主，512x512）：2019-03-08-stylegan-animefaces-network-02051-021980.pkl

百度网盘：https://pan.baidu.com/s/1KmauV9eEho9v6nINAdRmKA

提取码：dnbd

（4.3）一位中国研究生，从FFHQ人脸数据集中筛选了一些黄种人脸进行训练，生成了黄种人的人脸模型（1024x1204）：generator_yellow.pkl

百度网盘：https://pan.baidu.com/s/18cpaM6wJg4ozmwlFNY21kw

提取码：fx23

（5）修改生成人脸或少女动漫头像的文件

（5.1）进入stylegan-master目录（如：F:\AI\stylegan-master），创建cache目录，把人脸模型和动漫头像模型复制到cache目录下；

（5.2）修改pretrained_example.py文件，或者按下面的内容创建自己的pretrained_example001.py文件：

# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# This work is licensed under the Creative Commons Attribution-NonCommercial
# 4.0 International License. To view a copy of this license, visit
# http://creativecommons.org/licenses/by-nc/4.0/ or send a letter to
# Creative Commons, PO Box 1866, Mountain View, CA 94042, USA."""Minimal script for generating an image using pre-trained StyleGAN generator."""import os
import pickle
import numpy as np
import PIL.Image
import dnnlib
import dnnlib.tflib as tflib
import config
import glob
import randomPREFIX = 'Person'
#PREFIX = 'Animation'SEED = random.randint(0,18000)def main():# Initialize TensorFlow.tflib.init_tf()# Load pre-trained network.Model = './cache/karras2019stylegan-ffhq-1024x1024.pkl'#Model = './cache/generator_yellow.pkl'#Model = './cache/2019-03-08-stylegan-animefaces-network-02051-021980.pkl'model_file = glob.glob(Model)if len(model_file) == 1:model_file = open(model_file[0], "rb")else:raise Exception('Failed to find the model')_G, _D, Gs = pickle.load(model_file)# _G = Instantaneous snapshot of the generator. Mainly useful for resuming a previous training run.# _D = Instantaneous snapshot of the discriminator. Mainly useful for resuming a previous training run.# Gs = Long-term average of the generator. Yields higher-quality results than the instantaneous snapshot.# Print network details.Gs.print_layers()# Pick latent vector.rnd = np.random.RandomState(SEED)latents = rnd.randn(1, Gs.input_shape[1])# Generate image.fmt = dict(func=tflib.convert_images_to_uint8, nchw_to_nhwc=True)images = Gs.run(latents, None, truncation_psi=0.7, randomize_noise=True, output_transform=fmt)# Save image.os.makedirs(config.result_dir, exist_ok=True)save_name = PREFIX + '_' + str(random.getrandbits(64)) + '.png'save_path = os.path.join(config.result_dir, save_name)PIL.Image.fromarray(images[0], 'RGB').save(save_path)if __name__ == "__main__":main()

上面的Python文件是用来生成人脸的，如果需要生成动漫头像，只需要把

Model = './cache/karras2019stylegan-ffhq-1024x1024.pkl'

或：Model = './cache/generator_yellow.pkl'

修改为：

Model = './cache/2019-03-08-stylegan-animefaces-network-02051-021980.pkl'

即可。

（5.3）在stylegan-master目录下运行程序：python pretrained_example001.py

（5.4）到stylegan-master/results目录下查看生成的图像，That's all.

（5.5）也可以批量生成人脸或动漫头像，按下面的内容创建自己的pretrained_example002.py文件：

# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# This work is licensed under the Creative Commons Attribution-NonCommercial
# 4.0 International License. To view a copy of this license, visit
# http://creativecommons.org/licenses/by-nc/4.0/ or send a letter to
# Creative Commons, PO Box 1866, Mountain View, CA 94042, USA."""Minimal script for generating an image using pre-trained StyleGAN generator."""import os
import pickle
import numpy as np
import PIL.Image
import dnnlib
import dnnlib.tflib as tflib
import config
import glob
import randomPREFIX = 'Person'
#PREFIX = 'Animation'TIMES_LOOP = 100def main():# Initialize TensorFlow.tflib.init_tf()# Load pre-trained network.Model = './cache/generator_yellow.pkl'#Model = './cache/karras2019stylegan-ffhq-1024x1024.pkl'#Model = './cache/2019-03-08-stylegan-animefaces-network-02051-021980.pkl'model_file = glob.glob(Model)if len(model_file) == 1:model_file = open(model_file[0], "rb")else:raise Exception('Failed to find the model')_G, _D, Gs = pickle.load(model_file)# _G = Instantaneous snapshot of the generator. Mainly useful for resuming a previous training run.# _D = Instantaneous snapshot of the discriminator. Mainly useful for resuming a previous training run.# Gs = Long-term average of the generator. Yields higher-quality results than the instantaneous snapshot.# Print network details.Gs.print_layers()for i in range(TIMES_LOOP):# Pick latent vector.SEED = random.randint(0, 18000)rnd = np.random.RandomState(SEED)latents = rnd.randn(1, Gs.input_shape[1])# Generate image.fmt = dict(func=tflib.convert_images_to_uint8, nchw_to_nhwc=True)images = Gs.run(latents, None, truncation_psi=0.7, randomize_noise=True, output_transform=fmt)# Generate and Save image.os.makedirs(config.result_dir, exist_ok=True)save_name = PREFIX + '_' + str(random.getrandbits(64)) + '.png'save_path = os.path.join(config.result_dir, save_name)PIL.Image.fromarray(images[0], 'RGB').save(save_path)if __name__ == "__main__":main()

然后，在stylegan-master目录下运行程序：python pretrained_example002.py

生成的黄种人头像：