TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Partly: AI-Powered Art Generation from Your Photos

1 点作者 mickel超过 2 年前

1 comment

mickel超过 2 年前
I developed Partly, a side-project that uses Dreambooth and Stable Diffusion, built with Elixir and Phoenix LiveView.<p>Partly allows users to submit 10-15 photos of themselves or someone they know, which are then pre-processed and used to train a custom AI model with Dreambooth.<p>The model is then used to render 10 artworks each in 22 different styles (220 images in total), and the resulting images are post-processed, upscaled, and converted to webp format before being delivered to the user via email.<p>Despite the recent popularity of similar services, I believe the unique styles offered by Partly set it apart. Either way, it was a fun project and I&#x27;m happy with how it turned out.