Everything regulation | Flux Texture texture enhancer

Flux Texture Enhancement Has Arrived！

Welcome to Adjusting LAB. With the release of the FLUX model, the texture enhancer has also undergone significant updates. All materials are sourced from Midjourney, which can greatly enhance the details, texture, and clarity of the image, improving the stiff and rough texture of the FLUX base mold. The effect is extremely obvious when producing high-resolution images.

It is recommended to have a size of 1200 * 1600 or 1536 * 1536 to directly produce high-definition images and achieve the ultimate picture quality. Of course, 1024 * 1024 is also acceptable.

All training sets of this model are labeled using natural language, and standard natural language is used to describe the image you want. The model has a very high feedback rate for prompt words, and can completely achieve what you write. (For more details, please refer to all the display examples to easily understand the advantages of this model.)

Model Introduction：

触发词：Ultra HD quality（也可以不写）

推荐权重： 0.6

（0.2-0.8均可，可根据自己想要的细节程度调节权重；推荐0.6权重既可以增强细节和画质，又不会过多的改变原图）

基于flux底模的不同权重测试图：

（以下测试图均为带质感增强lora直出，均未二次放大处理，分辨率为1200*1680）

It can be clearly seen that the image produced by the FLUX bottom mold is stiff and blurry, and after adding LORA, the style and texture have been greatly improved. Influenced by the art style of Midjourney materials, the resulting images will change the rigid art style of the original FLUX base model, greatly enhancing the aesthetics of portraits, animals, materials, and scenes.

Side Notes:

Thank you very much to every friend in the Zhumeng community for their help during the model training process. I would like to express my special thanks to the head author of Lib, @ Dr-Dream, for their valuable opinions on the selection of training set materials and the model training process. Additionally, I would like to thank my fans for their support! Thank you!!