Fix InstanceNorm Shape docstring to require at least 3 dimensions

[ayaangazali]

Proposed changes

The InstanceNorm docstring Shape section says the input is (..., C), which reads like a 2D (N, C) input is allowed, but __call__ raises for anything with fewer than 3 dimensions:

import mlx.core as mx
import mlx.nn as nn

nn.InstanceNorm(4)(mx.zeros((2, 4)))
# ValueError: InstanceNorm expects inputs with at least 3 dimensions (N, ..., C) ...

Updated the Shape line to (N, ..., C) and noted the at-least-3-dimensions requirement, matching the error message and the reduction over the spatial axes. Same spirit as the earlier BatchNorm shape doc fix (#3782).

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)

(docstring-only, test_nn.py passes 69/69)

---

being honest: freshman here, Claude Code helps me find these mismatches but i reproduced the 3-dimension error above myself and read the reduction axes to make sure the constraint is real before opening this. corrections welcome :)