Make Batch.__eq__ work also for scalar values (0-dimensional arrays) #1182
Description
As first noticed here #1086 (comment), DeepDiff which is currently used for batch equality cannot handle 0-dimensional arrays.
My current proposal is to leverage np.atleast_1d and the newest Batch API's Batch.apply_values_transform to convert any 0-dimensional arrays to 1-d while keeping the others as they are (cf. #1086 (comment)).