Issue using invert_yaxis across a layout

[jlstevens]

Compare the following:

import holoviews
import numpy as np
import xarray as xr

ims = [xr.DataArray(np.random.rand(50,50), dims=['x','y'], 
                    coords={'x':np.arange(50), 'y':np.arange(50)}) for i in range(2)]

When viewing a single cell:

%%opts Image [invert_yaxis=True]
hv.Image(ims[0], kdims=['x', 'y'])

Versus a layout:

%%opts Image [invert_yaxis=True]
hv.Image(ims[0], kdims=['x', 'y']) + hv.Image(ims[1], kdims=['x', 'y'])

The layout requires panning upwards to see the data (which is there but not in the region initially shown) but only when invert_yaxis is specified.

This is using holoviews 1.11.0a8.