Subh775/sample-no-overfit download history
Subh775/sample-no-overfit is a token classification dataset on the Hugging Face Hub. In the last 30 days it was downloaded 23 times (9 in the last 7 days), and 446 times in total. It ranks #444,312 among datasets by monthly downloads.
sample-no-overfit A short-story dataset where each input is a non-overlapping context of 20 tokens, and the output is the same 20 tokens shifted by one position. This means no overlap between consecutive batches, reducing the risk of overfitting to the same text segments.
Open Subh775/sample-no-overfit on Hugging Face
Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.