creator: Liquidn2

🎈 What is Anima_LiquidMix? Anima_Liquid is an experimental merge model created to modify rendering behavior while preserving the original prompt understanding of Anima. It was built by integrating custom-made LoRA into multiple checkpoints in order to test whether texture and rendering characteristics can be controlled through merging.

This model can also be used as a base for LoRA training, since the internal structure remains compatible with the original Anima architecture after the merge. Because of this, it may be useful for testing LoRA behavior, training experiments, and rendering comparisons between different merged checkpoints.

🐲 Changes from Preview Base

Changed contrast behavior
Added variation in color and lighting
Enhanced texture rendering
Modified glow and effect rendering
Different tendencies when generating fantasy / furry / creature themed characters

🧪 Purpose

This model is not intended to reproduce a specific style. Instead, it was created as a test model to observe how merging and LoRA integration affect rendering behavior.

Because of this, output may vary significantly depending on:

prompt
sampler
seed
other generation settings

🍩 Recommended Use

Comparing different Anima-based checkpoints
Testing merge behavior
Testing LoRA-integrated models
LoRA training experiments
Generating fantasy / furry / insect / monster characters
Testing glow, effects, and high-contrast rendering

⚙️ Recommended Settings

Sampler: ER SDE / Euler a / DPM++
Steps: 20–30
CFG: 4–6
Resolution: 1024–1536

🧬 Requirements

This model is recommended for use with qwen_3_06b_base as the text encoder. For VAE, QwenimageVAE_liquid1087 is recommended for more stable rendering, color balance, and highlights.

📛 Notes

This is not an official model
This is an experimental merge created for testing purposes
Unexpected outputs or unstable behavior may occur

⭐ Feedback

Feedback and example generations are welcome.

Anima_LiquidMix - 3.8B_10138Dual