Datasets:

hazyresearch
/

ncm-tokenized-datasets

Dataset card Files Files and versions

ncm-tokenized-datasets

51.8 GB

Ctrl+K

Ctrl+K

2 contributors

History: 26 commits

kumboh's picture

wmt14: newstest2014 sources and untruncated references (3,003 sentences)

bb34d1c verified about 7 hours ago

full_owt
Add verified EOS document offsets for full OWT GPT-2 binaries 28 days ago
lm1b
lm1b: add train_bert_30522.bin (FLM-matched, bert-base-uncased, block 128) about 1 month ago
owt
owt: document offsets for the .bin files about 1 month ago
small_owt
Upload small_owt/valid_bpe_16384.pickle with huggingface_hub 6 months ago
sudoku
Add tokenized Sudoku datasets: easy, medium, hard 15 days ago
tinystories
tinystories: add train_bpe_8192_offsets.npy (story starts) 14 days ago
wikitext
Upload folder using huggingface_hub 10 months ago
wmt14
wmt14: newstest2014 sources and untruncated references (3,003 sentences) about 7 hours ago
xsum
xsum: add complete 11,334-document test set (passages.pickle) and untruncated references (reference_summaries.pickle) about 8 hours ago
.gitattributes

2.46 kB
initial commit 10 months ago