`deepcopy` doesn't deepcopy `.attrs['transform']` in `GeoDataFrame`

This fails:

    import geopandas as gpd
    from shapely.geometry import Point
    import copy

    # Create a GeoDataFrame
    gdf = gpd.GeoDataFrame({
        'geometry': [Point(1, 1), Point(2, 2)],
        'value': [1, 2]
    })

    # Add an attribute to the GeoDataFrame
    gdf.attrs['transform'] = {}

    # Copy the GeoDataFrame
    gdf2 = copy.deepcopy(gdf)

    # Print the IDs
    print(id(gdf.attrs['transform']))  # ID of copied attribute
    print(id(gdf2.attrs['transform']))  # ID of original attribute

    # Assertion
    assert gdf.attrs['transform'] is not gdf2.attrs['transform']

No source file or test is named. Start by running the provided GeoDataFrame and deepcopy reproduction, then trace how GeoDataFrame.attrs is copied. Done means the transform attribute in the copied GeoDataFrame is a distinct object while preserving its contents.