Finish I-JEPA

Experimental support for I-JEPA was added in #1273

We should do some refactoring and testing before fully releasing the model.

Todo

• #1367
• Verify that weights are correctly initialized and follow this
• Check if we can refactor IJEPAMaskCollator into a transform, comment.
• Move positional embedding functions (_get_1d_sincos_pos_embed_from_grid etc.) to lightly/models/utils and add reference to source
• Add missing docstrings or bring into correct format, check diff from #1273
• Add unit tests
• Finish pytorch lightning and distributed examples
• Add docs
• Add imagenet benchmark to lightly/benchmarks/imagenet/vit/ijepa.py
• #1712

Start by reviewing the experimental I-JEPA support from #1273 and the remaining checklist, including IJEPAMaskCollator and the positional embedding functions in lightly/models/utils. Inspect lightly/benchmarks/imagenet/vit/ijepa.py along with the referenced examples and diffs; done means the initialization, refactoring, tests, docstrings, examples, documentation, and benchmark items are complete.