GRAFFORD: A Benchmark Dataset for Testing the Knowledge of Object Affordances of Language and Vision Models