Fix broken all_sum and Group references in the data parallelism example - #3996

Proposed changes

The data parallelism example links to all_sum and Group with bare names, but that page has no currentmodule directive, so neither resolves and both render as plain text instead of links:

docs/src/examples/data_parallelism.rst:36: WARNING: py:func reference target not found: all_sum
docs/src/examples/data_parallelism.rst:36: WARNING: py:class reference target not found: Group

Both live under mlx.core.distributed, and the sibling page tensor_parallelism.rst already writes them fully qualified, so this brings the two pages in line. The other two references on the page were already fully qualified (mlx.nn.average_gradients, mlx.utils.tree_map), so these two were the odd ones out.

After the change the page builds with no unresolved references.

Checklist

Put an x in the boxes that apply.

• I have read the CONTRIBUTING document
• I have run pre-commit run --all-files to format my code / installed pre-commit prior to committing changes
• I have added tests that prove my fix is effective or that my feature works
• I have updated the necessary documentation (if needed)

(docs only; verified by building the docs with -n before and after)

usual note on process: freshman contributor, i lean on Claude Code for the sweeping, but i built the docs in nitpick mode myself, read the warnings, and checked how the neighbouring tensor parallelism page writes the same two references before picking the fix.