Multichannel input (TIF files)

Is it possible to use images with more than three channels? For instance, the images typically used by this model are in RGB format. However, could it also handle TIF images that include RGB + NIR (near-infrared) channels? If so, which files would need to be modified?

The issue does not name any files, tests, or entry points. Start by tracing how image channels are loaded and how model inputs are defined, then determine the affected training and inference paths. Done means establishing whether RGB+NIR TIFF input is supported and documenting or implementing the required changes.

Written by the indexing model from the issue text.