Windows 剪贴板增强工具，把每次复制的内容存进数据库，之后随时翻出来重新粘贴，不用再反复复制。
LTX-Video 是 Lightricks 开源的视频生成模型，支持文生视频和图生视频，基于 DiT 扩散架构，可快速生成高质量短视频。
SANA 是英伟达开源的高效扩散生成模型与配套训练推理引擎，支持文生图、文生视频和世界模型，用线性注意力和 NVFP4 量化在低显存下实现快速生成。
开源视频生成模型，用高效架构和训练方案降低高质量视频制作门槛，让普通开发者也能训练和部署文生视频模型。
腾讯混元开源的大规模视频生成框架，基于扩散 Transformer 把文本提示直接生成连贯视频，适合做文生视频研究和应用落地。
…ection synchronization overhead. - Strictly align DiT weight initialization to the [original DiT implementation](https://github.com/facebookresearch/dit). Fixed - Fix NaN
…unfiled] OOB write in symbol-mode Floyd-Steinberg dithering. [unfiled] CLI: Incomplete sanitization of filenames printed to terminal. [unfiled] CLI: Integer overflow in X
…l by @xbeta-zhang in https://github.com/sabrogden/Ditto/pull/906 updated french translation by @sabrogden in https://github.com/sabrogden/Ditto/pull/911 Added back in loc
…ler by @sabrogden in https://github.com/sabrogden/Ditto/pull/905 Full Changelog : https://github.com/sabrogden/Ditto/compare/3.25.73.0...3.25.76.0
…cript by @bovirus in https://github.com/sabrogden/Ditto/pull/467 delete old italian language / Add new Italian language by @bovirus in https://github.com/sabrogden/Ditto/
…int by @sabrogden in https://github.com/sabrogden/Ditto/pull/971 Added special paste option to paste a new Guid by @sabrogden in https://github.com/sabrogden/Ditto/pull/1