The saved dataset is saved in multiple file "shards". By default, the dataset output is divided to shards within a round-robin vogue but personalized sharding is usually specified by using the shard_func function. For instance, It can save you the dataset to using only one shard as follows:Key activities of SCF could be divided into 3 regions: one