One pose I always struggled to generate with anima models was placing the character’s feet behind the head. Most of the time, the model would put the feet beside the head or in front of the face instead, making this pose surprisingly difficult to generate consistently.

So I created this LoRA specifically to improve the success rate of that pose. Its main purpose is to guide the model toward positioning the feet correctly behind the head, rather than changing the overall art style.

For the LoRA weight, I recommend using 0.7 to 0.85.

If you are using a dual-sampling Anima workflow, I recommend applying this LoRA only during the first sampling stage. Do not apply it again in the second sampler, as this may over-strengthen the pose or negatively affect the final result or increase the toes number.

I hope this LoRA makes the pose much easier and more consistent to generate.