Same as previous one. Just slightly modified. Use V1 for now. Trying this out. Not sure if it would work. Very experimental stuff. Using the techniques of Bias-Free LoRA adapted for Anima. By making the merged LoRAs not influence the first few blocks. This model is meant to be used alone, without LoRAs! If you are using LoRA/spells, try using them at low strengths. ~0.1 - 0.5