leafyeexyz / painterleaf

A Cute WebApp. Text2Image, Image2Image, Image2Text, Save or Star Images, Multiple Models, Chinese&English Prompt Suport. 一个可爱的 AI 绘画 WebApp, 支持文生图, 图生图, 图片收藏和下载, 多模型, 中英双语提示词. 图生文

简体中文 | 正體中文 | English

PainterLeaf

• Text-to-image, supports multiple models
• Image-to-text, convert local images to prompts
• Also supports image-to-image
• Prompt supports Chinese and English (Chinese will automatically call AI translation)
• Front-end and back-end separation, front-end based on React, back-end based on Hono, see this project
• API provided by CloudflareAI and HuggingFace
• Internationalization support, currently supports 简体中文, 正體中文, and English

Usage

Deploy Server

See this project

Modify src/config.json

{
  "SERVER": "https://api.xxx.workers.dev", // Your server address
}

Install Bun

Please refer to Bun.sh. Or simply run npm i -g bun

Install dependencies

bun i

Local run

bun run dev

Build

bun run build

painterleaf's Issues

Apply React Compiler

要不做个图生图的模式

加一个从图像获取提示词的功能

给每张图片添加一个下载按钮

UI 加入中英双语支持

右上角添加一个按钮来显示关于页对话框

CUDA Out Of Memory Errors

Hello, nice work here.

I'm running into some memory management issues when using this, however. It's not happening on every single run, and I haven't identified any pattern that causes it yet, so currently it remains an intermittent error. But I've run into it 4 times now, out of a total of 14 generation runs.

Here's the error text that appears in the modal dialog:

生成失败
服务端返回错误信息: {"error":"unknown error","warnings":["CUDA out of memory. Tried to allocate 1024.00 MiB (GPU 0; 14.75 GiB total capacity; 8.46 GiB already allocated; 985.06 MiB free; 8.99 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF","There was an inference error: CUDA out of memory. Tried to allocate 1024.00 MiB (GPU 0; 14.75 GiB total capacity; 8.46 GiB already allocated; 985.06 MiB free; 8.99 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF"]}

🤷🏼‍♀️

加个自动主题切换

为了不增加更多的按钮，始终跟随系统设置

加个图片云分享功能

主要解决不同设备的收藏图片的传递问题。考虑通过 CloudflareKV 实现，采用一定的认证机制来控制能否上传。后端新增三个接口：获取是否在云端，Get，参数为hash，下载图片，Get，参数为hash，上传图片，Post，参数为图片JSON对象和密码

使用原生 IndexedDB API 以试图修复Safari 上的 Bug

重构为纯 React 项目