A depth estimation model using a SwinV2-Tiny backbone for monocular depth prediction at 256x256 resolution.