腾讯推出全新开源模型混元DiT：6G小显存版本支持Kohya训练，引领AI生成图像新纪元
近日，腾讯公司推出了一项重大更新——混元DiT模型。这款广受好评的AI生成图像模型迎来全新升级，旨在帮助开发者更轻松地进行图像创作与模型训练。这一更新为整个行业带来了新的活力和革命性的进步。
新推出的版本包括了诸多亮眼的新特性。首先，腾讯宣布开源小显存版本，仅需6G显存即可运行，极大减轻了开发者的硬件负担，对使用个人电脑本地部署的开发者十分友好。同时，这一版本还成功适配了LoRA、ControlNet等插件至Diffusers库，大大提高了开发者的工作效率。特别值得一提的是，新版本的混元DiT还支持Kohya图形化界面，让开发者可以低门槛地训练个性化LoRA模型，极大地降低了使用门槛。此外，混元DiT模型已升级至1.2版本，不仅在图片质感上有所提升，而且在构图方面也展现出更加出色的表现。这对于图像创作领域无疑是一次重大突破。这一更新将为图像生成和人工智能创作领域带来新的活力与挑战。目前该模型的更新进展获得了广大用户和业内人士的普遍关注与高度期待。今后随着技术进一步发展，我们期待看到更多这样的创新和突破出现在人工智能领域中。
英语如下：
News Title: “Tencent MixTu DiT Upgrade: 6G Small Texture Memory Open Source, Kohya Support Assists Developers in Training New Era”
Keywords: Tencent MixTu, Upgrade, Support Kohya Training
News Content:
Tencent recently launched a significant update – the MixTu DiT model. This highly praised AI image generation model has undergone a new upgrade to assist developers in image creation and model training, ushering in a new era of AI image generation.
The newly released version includes many impressive new features. Firstly, Tencent announced the open source of a low-memory version that can run with just 6G of graphics memory, greatly reducing the hardware burden on developers and being friendly to developers who use personal computers for local deployment. Meanwhile, this version has also successfully integrated plugins such as LoRA and ControlNet into the Diffusers library, greatly improving developers’ efficiency. Notably, the new MixTu DiT also supports the Kohya graphical interface, enabling developers to easily train personalized LoRA models. In addition, the MixTu DiT model has been upgraded to version 1.2, which not only improves image quality but also demonstrates superior performance in composition. This is undoubtedly a major breakthrough in the field of image creation. This update will bring new vitality and challenges to the image generation and AI creation fields. Currently, the update progress of this model has received widespread attention and high expectations from users and industry insiders. As technology continues to develop, we look forward to seeing more such innovations and breakthroughs in the field of artificial intelligence.
【来源】https://mp.weixin.qq.com/s/SoetAIDZAcebpOvAafJRcA
Views: 3