Datasets:

devingulliver
/

dolma-v1_6-sample

Modalities:

Formats:

Languages:

Size:

Libraries:

Dataset card Data Studio Files Files and versions Community

devingulliver commited on Apr 21, 2024

Commit

6b03e4b

·

verified ·

1 Parent(s): 2cb2597

Update README.md

Files changed (1) hide show

README.md +2 -0

README.md CHANGED Viewed

@@ -19,3 +19,5 @@ configs:
   - split: train
     path: data/train-*
 ---

   - split: train
     path: data/train-*
 ---
+Taken directly from [allenai/dolma](https://huggingface.co/datasets/allenai/dolma). Roughly 10B tokens. Useful for small scale testing or maybe continued pretraining.