| license: apache-2.0 | |
| description: | |
| - the token indexes for the wikitext-103 dataset. This is a repackaging of the data from "AMead10/wikitext-103-v1-gtp2-tokenized". | |
| - In this dataset, each split is a torch.save() file containing a single tensor of the token indexes. |