This dataset is the same as the 'test.ft.txt.bz2' file made available at https://www.kaggle.com/datasets/bittlingmayer/amazonreviews.
There are only 2 differences:
1. the file format
2. the labels: I`ve changed __label__1 for 'bad' and __label__2 for 'good'.
