CN Workflow - Dragon Boat Festival Animal Zongzi Scene ACG6 lightflow Hyper

File name: CN Workflow - Dragon Boat Festival Animal Zongzi Scene ACG6-lightflow.flow

----------------------------------------------------------------------

【This workflow is suitable for AiARTiST exclusive ACG Playground SDXL V6 version】

Direct link:  https://www.liblib.art/modelinfo/4bb107737a1349f7a053bbf618e46977

LoRA model link: No need for LoRA (other SDXL base models can also run, CN control still works, but the style may not be consistent with the example image)

CN controller model: https://www.liblib.art/modelinfo/d1d7104c109948aead3d7c7b29e13aaf

CN controller model storage path: webui\models\ControlNet\

CN preprocessor method: softedge_teed (optional, you can directly drag the control graph from the example image into CN)

AiARTiST model list: https://www.liblib.art/search?keyword=AiARTiST

Where,

Free models account for 55%, open-source models account for 95%

Member models account for 40%, closed-source models account for 5%

----------------------------------------------------------------------

----------------------------------------------------------------------

Instructions: Download the Flow file to your local machine, drag it into the LightFlow plugin box, and it will automatically fill in parameters and control graphs

The lightflow plugin is included in common webui integration versions; if not, you can download and install it yourself (FORGE needs to be installed separately):

https://fvkij7wuqx9.feishu.cn/docx/HgZndihraotmmzxFni7cFZISnvb

Thanks to MistoLine development team Control Every Line!

Citation

@misc{
    title={Adding Conditional Control to Text-to-Image Diffusion Models},
    author={Lvmin Zhang, Anyi Rao, Maneesh Agrawala},
    year={2023},
    eprint={2302.05543},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}

Thanks to lightflow.ai development team

LightFlow Vision

Vision: Lower the threshold of SD usage and achieve AI inclusivity

Goal: Promote widespread use of SD technology, making it easier to convey AI drawing abilities, reducing learning costs, and increasing efficiency through plugins

----------------------------------------------------------------------

Preprocess image (can be directly dragged into ControlNet input box), refreshing the model to select MistoLine will take effect.

Flow parameters are adapted for Hyper base models, manual changes are required when switching to other base models.

My common configurations:

Clip Skip 1 / CFG 1 / Steps 10 / Sampling method (Sampler) Euler a / VAE Auto

Default: 768 x 1536, upscale high-def 1.5x, redraw high-def magnitude 0.5, redraw method Latent

Since the final composition size is sufficient, ADetailer repair is not needed.

Recommended configuration without HYPER acceleration:

Clip Skip 2 / CFG 7 / Steps 30 / Sampling method DPM++ 2M Karras / VAE Auto

Default: 640 x 1280, upscale high-def 1.5x, redraw high-def magnitude 0.5, redraw method Latent

----------------------------------------------------------------------

Hyper-SD is one of the latest diffusion model acceleration technologies, lossless high-def acceleration, CN is well adapted

After testing with multiple base models, Hyper-SD can derive more image information than before acceleration!

Strongly recommend users with 8GB of VRAM and MAC to use the Hyper version, directly outputting 2048 HD images in about 10 seconds

----------------------------------------------------------------------

Lib members can enjoy online accelerated generation, conveniently used in combination with CADS series LoRA.

Model list: https://www.liblib.art/search?keyword=AiARTiST

Commercial models:

AiARTiST-CADS2.0 XL Business Advertisement Assistance System (Enterprise Customization)

https://www.liblib.art/modelinfo/a56ebacdba7d4e30b97bb124bc3fc28f <<Click for direct access

----------------------------------------------------------------------

Model usage declaration:

1. You may not host this model and its derivative versions (such as fused model versions) on websites/applications designed to earn revenue or donations.

2. You may not directly sell this model and its derivative versions (such as fused model versions) unless you have made a sufficient degree of manual modifications to the model so that it can be fully deemed your personal work in a legal sense. If you violate this clause, you will be personally responsible for all legal consequences, and I will not be held responsible.

3. You may not use the model to intentionally create or share illegal or harmful content for dissemination and output, please adhere to good moral standards, and use the model for positive purposes.

----------------------------------------------------------------------

For testing issues, please leave a message, for business partnerships check personal homepage +V Zyuan980

AiARTiST | Metaverse China·Shandong Digital Human/AiGC/Metaspace Direction

Be a good tool for artists Contact information: https://afdian.net/a/AiARTiST-lib