此项目只能随机生成图片,不能以图生图
先把源码克隆到本地,https://github.com/NVlabs/stylegan3
进入项目文件夹,激活虚拟环境
conda env create -f environment.yml
conda activate stylegan3
访问https://pytorch.org/get-started/locally/,根据自己的环境选择
在虚拟环境执行
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
生成图片,用官方的模型
# Generate an image using pre-trained AFHQv2 model ("Ours" in Figure 1, left).
python gen_images.py --outdir=out --trunc=1 --seeds=2 --network=https://api.ngc.nvidia.com/v2/models/nvidia/research/stylegan3/versions/1/files/stylegan3-r-afhqv2-512x512.pkl
生成视频,用官方的模型,这个电脑不行的话,会有点慢
# Render a 4x2 grid of interpolations for seeds 0 through 31.
python gen_video.py --output=lerp.mp4 --trunc=1 --seeds=0-31 --grid=4x2 --network=https://api.ngc.nvidia.com/v2/models/nvidia/research/stylegan3/versions/1/files/stylegan3-r-afhqv2-512x512.pkl
结果
若生成图片报错
OMP: Error #15: Initializing libiomp5md.dll, but found libiomp5md.dll already initialized. OMP: Hint This means that multiple copies of the OpenMP runtime have been linked into the program. That is dangerous, since it can degrade performance or cause inco