license: apache-2.0
This data set is created by randomly sampling 1M documents from the large supervised proportional mixture from the T5 repository.
The code to produce this sampled dataset can be found here.
