Possible issue with batched G_ tensor in QPFunctionFn_infeas.forward



In `bindings/python/proxsuite/torch/qplayer.py`:

```cpp
class QPFunctionFn_infeas(Function):
    @staticmethod
    def forward(ctx, Q_, p_, A_, b_, G_, l_, u_):
         n_in, nz = G_.size() # <- can't G be batched? If so, this will cause an issue.
```


If `G_` can be batched (e.g. shape `(batch_size, n_in, nz)`), this line will break. Might need a shape check or to handle batching explicitly.