Align3R: Aligned Monocular Depth Estimation for Dynamic Videos | Read Paper on Bytez