F.1-Dream Mirror | Cubism Fashion

This Lora is an upgraded version of Dream Mirror | Cubism Flux, with some minor fixes on the original dataset, but it is basically the original dataset of SD1.5 version.

The training materials mainly come from contemporary cubist master Georgy Kurasov, a painter from St. Petersburg, Russia, also known as the "new" constructivist painter. He inherited the ideas of constructivism, piecing together the characters in his writing with geometric fragments and often placing them in irregular geometric spaces resembling kaleidoscope fragments. At the same time, he used vivid colors, which formed a huge contrast with the cold and stern characters in his writing.

For details, please refer to Dream Mirror | Cubism SD1.5 version:

https://www.shakker.ai/zh-TW/modelinfo/00399d3f48ed4960a77df96844788c8c/Dream-Mirror-Cubism

Recommended size: 768 * 10241024 * 10241152 * 1536 (ComfyUI online workflow has high-definition enlarged nodes, recommended 1.5-2 times for better detail presentation)

Recommended Steps: Starting from 20

Trigger word: litizhuyi-bb

Example prompt words: litizhuyi-bb, A Surrealist Portrait of a Woman with an Umbrella and Her Dog, Set in a Vibrant Room

Note: Flux suggests using natural language for the entire model to enhance the logical understanding of semantics, such as orientation, structure, color, detailed description of specified objects, and description of relationships between objects.

The charm of flux is infinite, and the rest depends on the calories burned by everyone. Take advantage of the heat and connect three times with one click. See you in the picture area~

If deployed locally, you can download any image after the second one and drag it in to reproduce the workflow.

Online workflow recommendation: (Replace the main Lora with this Lora Dan)

1-ComfyUI version of Libuwen Shengtu online experience address: https://www.liblib.art/modelinfo/ef7eb3116045485f831985c64034e2ba 2. We have specially built an online workflow for the conversion of images from production to drawing. You can try it out, and it is already the formula I have debugged for the e****r. The effect is quite good: https://www.liblib.art/modelinfo/c9b347e5c5914c7db208fe1f6da7ddb9

Note: Flux suggests using natural language for the entire model to enhance the logical understanding of semantics, such as orientation, structure, color, detailed description of specified objects, and description of relationships between objects.

The charm of flux is infinite, and the rest depends on the calories burned by everyone. Take advantage of the heat and connect three times with one click. See you in the picture area~

If deployed locally, you can download any image after the second one and drag it in to reproduce the workflow.

Online workflow recommendation: (Replace the main Lora with this Lora Dan)

1-ComfyUI version of Libuwen Shengtu online experience address: https://www.liblib.art/modelinfo/ef7eb3116045485f831985c64034e2ba 2. We have specially built an online workflow for the conversion of images from production to drawing. You can try it out, and it is already the formula I have debugged for the e****r. The effect is quite good: https://www.liblib.art/modelinfo/c9b347e5c5914c7db208fe1f6da7ddb9