TAGA: Self-supervised Learning for Template-free Animatable Gaussian Articulated ModelShare on Twitter Facebook LinkedIn Previous Next