[BUG] DataFrame.columns returns a pd.Index not a cudf.Index

[Matt711]

Describe the bug
When accessing the columns of a cudf.DataFrame initialized with with a columns=cudf.Index(...), a pandas index is returned.
Steps/Code to reproduce bug

import cudf

df = cudf.DataFrame(columns=["foo"])

print(df)
print(type(df.columns))

Empty DataFrame
Columns: [foo]
Index: []
<class 'pandas.core.indexes.base.Index'>

Expected behavior
return a cudf.Index

[vyasr]

#14012 (comment)