ProsePainter是一款基于机器学习的有趣工具

python nodejs

它可以通过文字描述来生成图像。遵守GPL3.0开源协议。

程序包含一个python服务器用来计算，一个使用nodejs服务器的网页端界面 .用户只需要描述想画的东西和在那个区域，ProsePainter就会根据机器学习的实时计算生成图片。

安装：
安装Python服务器
建议使用python3.7 以及 torch 1.7.1+cu110 以上版本(这里下载：https://pytorch.org/get-started/previous-versions/).

建立 conda 环境，取名 prosepaint.

conda env create -f server/env-server.yml

运行 Python 服务器

运行命令启动 uvicorn 服务器.

export PYTHONPATH=.; python server/server_deploy.py

cd client
npm install

运行NodeJS 服务器

cd client
npm run dev

浏览器打开 http://localhost:8003/ 即可使用。

源码：https://github.com/Morphogens/ProsePainter

-------

ProsePainter combines direct digital painting with real-time guided machine-learning based image optimization. Simply state what you want and draw the region where you want it. Tux, the Linux mascot

Setup

The app consists of a python server which does the computations and a web based interface using a nodejs server.

Install Python server

Due to the use of CLIP, it is recommended to use python3.7 and torch 1.7.1+cu110 (available here).

With the following command you can set up a conda environment named prosepaint where all the dependencies will be installed.

conda env create -f server/env-server.yml

Run Python server

The following command will launch the uvicorn server.

export PYTHONPATH=.; python server/server_deploy.py

Install NodeJS server

cd client
npm install

Run NodeJS server

cd client
npm run dev

Open http://localhost:8003/ in your web browser.

from 

https://github.com/Morphogens/ProsePainter