Femix_HassakuXL

FeMix_HassakuXL is a series of Stable Diffusion XL models designed for generating anime-style images, employing a unique blend strategy to combine different aesthetic styles. The series ranges from highly stylized and experimental blends to stable, high-quality models suitable for general use. FeMix_HassakuXL merges two distinct models and concepts of each blend, to achieve various output depending on users needs.

Recommended Prompts & Settings

Quality Prompts: masterpiece, best quality, detailed
Negative Prompts: worst quality, bad quality, sketch. Adding "signature" can help remove unintended watermarks.
High-Quality faces: Put nose into the negative prompt for smooth faces, if you feel that Style A gets too prominent.
Sampling Method: DPM++ 2M Karras. Some users experience good results with Euler A and different samplers due to having weaker style implementation to improve image creation.
General Settings: Resolution 832x1216. Try varying different resolutions for unique image output.

General Recommendations:

Avoid overloading the model with lengthy prompts. Its versatility is also its weakness, as too much information can lead to unpredictable results.

Excluded metadata and franchise tags, so please do not use them.

Negative Recommendations:

* Avoid using disruptive features: floating texts, logos, watermarks, speak bubbles, signatures. If those results occur, add "signature" into the negative prompt.

* Excluded metadata and franchise tags, do not use them.

Recommended Resolutions (SDXL):

* 1024 x 1024

* 1152 x 896

* 896 x 1152

* 1216 x 832

* 832 x 1216 (most recommended)

* 1344 x 768

* 768 x 1344

* 1536 x 640

* 640 x 1536

License Info:

This model merges aspects from ANIMAGINE XL 3.0 (pre-V1) and Illustrious-XL & WAI-NSFW-illustrious-SDXL (V1 onwards):

Fair AI Public License 1.0-SD.

Description

Model Description: Style A+B (50/50 Blend)

This model is a fusion of Style A and Style B, offering a unique blend of characteristics from both. It aims to strike a balance between the anime-influenced aesthetic of Style A and the experimental style of Style B.

Key Features:

Brightness and Anime Influence: Inherits the brighter overall image appearance and the tendency towards a more anime-style aesthetic from Style A. Expect fewer highly-detailed/realistic noses and faces (especially without prompting), leaning towards a smoother, anime-inspired look.
LORA Compatibility: Maintains good compatibility with LoRA models thanks to the Style A influence. LoRAs should be able to significantly impact the style and details of your generations, including pony LoRAs.
Interesting Style with Potential Flaws: Incorporates the interesting and potentially flawed style of Style B. Be aware of potential artifacts (especially around the eyes) and possible composition issues. Highres fix or Adetailer is recommended to mitigate these flaws.
Balance of Styles: Aims to provide a middle ground between the cinematic lighting potential of Style A and the unique style of Style B. Results may vary, and experimentation is encouraged.

Recommendations:

Use highres fix or Adetailer to address potential artifacts and composition issues, inherited from Style B.
Experiment with different samplers to find one that minimizes artifact susceptibility.
Leverage LoRA models to further customize the style and details of your generations.
Consider using artist tags to refine the desired style, as the Style A base retains a good level of tag responsiveness.

Please Note: This is a merged model, and results may be unpredictable. Experimentation is key to discovering its full potential and understanding its limitations. The future development of a refined "Style B" variant depends on the popularity and feedback received for this merged model.