(Exclusive) Absolute Field #New 3D

Developer Notes：

Very welcome to visit my model site. Here, I would like to explain that this model is a derivative work created by me based on Animagine's base model (with various self-refining NTC optimizations). The current character style leans more towards realism, with optimizations in eyes, facial features, skin texture, and lighting in this version. In the new version, a brand new structure has been incorporated into the base set, making it easier for you to highlight certain areas.The visual impact of focal length on the image has also been greatly enhanced, leading to a more profound spatial understanding of the model, showcasing XL's features very well, allowing for everyone's boundless imagination!! PS: If there are more playable techniques, everyone must help UP by commenting and providing feedback for mutual learning and growth, Thank you all!!

Update Notes:

New version AbsoluteField#New3D_Level UP has undergone super optimization. In the new version, the framework has been optimized with GPT4 Vision for prompt word optimization, making character backgrounds more realistic and easier to understand, enhancing shadow effects, using Hyper for model acceleration, transforming into a lightweight XL with low memory consumption. Note: Set at 1 CFG, generate 8 steps images. More light source visualizations, Chinese style emphasis, neon enhancement, increased streamlined feel, tested stable output with Euler a, marking (DPM++ 2M characters slightly overfitting may cause blurring).

Comparison of New and Old Versions using the Same Seed:

Model Details:

Base Model: AnimagineXL3.1   Refined Official Model: AdapterPLUS

Embedded multiple self-refined NTC optimization sliders (NTC increased from 5 to 11), fundamentally changing the art style.

NTC weights: 2/1/0.8/0.8/0.5/0.5/0.3/0.2/0.1/-0.5/-1/ (If required, subsequent NTC sliders will be gradually made public based on demand)

Model Dimensions: #265# Model Accuracy: bf16

Embedded VAE: sdxl_0.9_vae.safetensors

Character List (referencing Qiu Qilin shared document by DaDa): https://docs.qq.com/sheet/DWGxXbEZxdmtXSERT

•Default Keywords: masterpiece, best quality, reality, realistic, Cinematic Lighting, 1girl

•Recommended Resolutions: 768*1365, 1024*1024, 1024*1280, etc.

•Recommended Steps: 8 or more (suggested 8)

•Recommended Sampling: Euler a

•CFG Scale: 1

•Focus on Tags: 1girl / 1boy, character name, others

•High-quality Hints like: masterpiece, best quality,

•Negative Hints like: nsfw, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name

Notes:

If you find a beautiful image, please give it a like and share more images! If you have any questions, feel free to leave a comment for me, I will accelerate the pace of updates! Thank you for your support!

Warning:

This model is strictly prohibited for any commercial purposes, only for educational and communication use. You cannot directly sell images generated by this model without my permission, you are not allowed to repost the model on other websites or sell this model without my consent, no fusion of this model, respect others, do not use this model to generate images that violate public order and morals or insult others. You will be personally responsible for any legal consequences arising from this, I will not be held responsible!!!