User Manual (Must Read):

This model is a Korean comic style model, which uses original materials to train the comic worldview and split format comics. It can generate any comic-style scenes/characters/vehicles, etc. This model can be used for novel promotions. With the help of WellCap, this model was trained using fully manual natural language labeling, and its prompt understanding ability is very strong. It is trained based on SDXL base 1.0, and the recommended weight ratio is 0.8~1.0.

The detailed usage of the model is as follows:

The basic trigger word is: Dream Comic Style (Please write this prompt at the beginning)

For split format comics, please add the following after the trigger word: Split Image Format

To increase the intensity of the comic style, please add the following after the trigger word: Comic Strip

This model is an art style model without fixed prompts. Please describe the scenes you have in mind using standard NLP natural language format. It can 100% restore the prompts, so feel free to unleash your imagination and create your own comic world!

This model is recommended to use After Detailer to depict facial details more clearly.

Note: Add your favorite color before the trigger word to change the model elements' colors. It is recommended to download and use local generation for higher resolution, which can provide more details! The recommended resolution should not be lower than 1024x1024! For online resolution, make sure it is not lower than 1024x1024!