功能概述
输入一段英文文本，和一张图片，生成符合文本描述和图片中人物姿态的人物图片
输入示例：
文本描述"：a woman standing by the sea"
图片：)
输出示例：
输出为：
环境准备
从github下载代码：
git clone https://github.com/ZcyMonkey/HumanSD_for_modelscope.git
进入文件夹
cd HumanSD_for_modelscope
配置HumanSD环境
conda install pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cudatoolkit=11.3 -c pytorch
pip install -r requirements.txt
配置MMpose环境
按照 MMPose 配置MMPose环境，推荐安装0.29.0版本的MMPose
运行代码
在根目录下运行 test_HumanSD.py 即可
示例代码
from modelscope.models import Model
from modelscope.pipelines import pipeline
import ms_wrapper
import os
model = "damo/Pose-driven-image-generation-HumanSD"
image_dir = "fff1daa9ef182b86387d802dba686426e7feb396_19692096.jpg"
prompt = "a woman standing by the sea"
neg_prompt = "lmonochrome, lowres, bad anatomy, worst quality, low quality"
inference = pipeline('Pose-driven-image-generation', model = model,model_revision="v1.0.1")
inputs = {"image_dir":image_dir,"prompt":prompt,"neg_prompt": neg_prompt,"sample_steps":30,"seed":None,"guidance_scale":10.0,"num_samples":2 }
output = inference(input=inputs)
output_dir = 'test'
if not os.path.exists(output_dir):
os.mkdir(output_dir)
for i in range(len(output)):
output[i].save(os.path.join(output_dir,str(i).zfill(4)+'.jpg'))
其中
inputs = {"image_dir":image_dir,"prompt":prompt,"neg_prompt": neg_prompt,"sample_steps":30,"seed":None,"guidance_scale":10.0,"num_samples":2 }
为输入参数，可以根据需求更改。
引用
@article{ju2023humansd,
title={Human{SD}: A Native Skeleton-Guided Diffusion Model for Human Image Generation},
author={Ju, Xuan and Zeng, Ailing and Zhao, Chenchen and Wang, Jianan and Zhang, Lei and Xu, Qiang},
booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
year={2023}
}
```
评论