I was honestly quite surprised to discover just how difficult it is for the Anima base model to render a anatomically correct Asian squat or deep squat pose for its characters. To bridge this gap, I spent some time training this dedicated LoRA specifically designed to assist and stabilize the deep squat action, making it reliable for your generations. Trained on a curated dataset of 35 diverse images spanning different artistic styles, this LoRA effectively eliminates those annoying, floating tiptoes, forcing the soles of the feet to stay completely flat and firmly attached to the ground.

For the absolute best results, I highly recommend keeping the model strength range between 0.8 and 1.0. Honestly, while it is technically possible to hit an Asian squat in vanilla Anima without any external help, your chances of getting a natural, anatomically accurate deep squat pose are incredibly rare—I’d say about 1 out of 10 at best. This LoRA completely changes those odds. Please feel free to push the limits, explore your creativity, and most importantly, have fun creating!