Remove grouping example

- The plan is to add it back in another PR

Author: ValerianRey

docs/source/examples/grouping.rst

@@ -29,9 +29,6 @@ This section contains some usage examples for TorchJD.
 - :doc:`PyTorch Lightning Integration <lightning_integration>` showcases how to combine
   TorchJD with PyTorch Lightning, by providing an example implementation of a multi-task
   ``LightningModule`` optimized by Jacobian descent.
-- :doc:`Grouping <grouping>` shows how to apply an aggregator independently per parameter group
-  (e.g. per layer), so that conflict resolution happens at a finer granularity than the full
-  shared parameter vector.
 - :doc:`Automatic Mixed Precision <amp>` shows how to combine mixed precision training with TorchJD.
 
 .. toctree::
@@ -46,4 +43,3 @@ This section contains some usage examples for TorchJD.
     monitoring.rst
     lightning_integration.rst
     amp.rst
-    grouping.rst

docs/source/examples/index.rst

@@ -41,14 +41,6 @@ class GradVac(GramianWeightedAggregator):
         For each task :math:`i`, the order of other tasks :math:`j` is shuffled independently
         using the global PyTorch RNG (``torch.randperm``). Seed it with ``torch.manual_seed`` if
         you need reproducibility.
-
-    .. note::
-        To apply GradVac with per-layer or per-parameter-group granularity, create a separate
-        :class:`GradVac` instance for each group and call
-        :func:`~torchjd.autojac.jac_to_grad` once per group after
-        :func:`~torchjd.autojac.mtl_backward`. Each instance maintains its own EMA state,
-        matching the per-block targets :math:`\hat{\phi}_{ijk}` from the original paper. See
-        the :doc:`Grouping </examples/grouping>` example for details.
     """
 
     def __init__(self, beta: float = 0.5, eps: float = 1e-8) -> None: