Change the DEFAULT_CONFIG_NAME

Convert the dataset to streaming mode.

Update the scenarios and intents.

Add the capability to get the entire corpora using the 'all' subset.

Add the BIO format for the named entities.

Update README

Add download_and_extract

Fix the issue with en-US like in the documentation

Add data loader for Amazon Massive corpus