Impermanence_cosplay_3D_anime_style F.1

Introduction to the selected LORA model: This model is trained based on the F.1-dev-fp8 base model and features a 3D anime style.

Recommendation: Strongly recommend using the Flux image generation workflow, one image with 3 compute powers, what are you waiting for?

Workflow address: https://www.liblib.art/modelinfo/9c81aa8555d2425490f7ecdf028f9abd?from=personal_page

Generation suggestion: F.1 basic algorithm model - runnable on the Lyu Bu online platform,

Sampling method: Euler

LORA weight: Recommended 0.8-1.0

Recommended size: This model is trained on a 1024*1536 training image set.

Iterations: 15-30, it was found that even 10 steps can produce good results.

Prompt word correlation coefficient: 3.5.

HD restoration: R-ESRGAN_4x+ Anime6B

Creative summary: Try training a LORA model with a new FLUX model to create a 3D anime style LORA, any unexpected surprises for anime characters under the FLUX large model? It is recommended to use the F.1 model of the workflow for image generation, which can save a lot of compute power.

Trigger word: cg art