
DatasetRank #498 of 836 155 likes
minipile
by JeanKaddour (personal account)
The MiniPile Challenge for Data-Efficient Language Models Dataset Summary MiniPile is a 6GB subset of the deduplicated The Pile corpus. To curate MiniPile, we perform a simple, three-step data filtering process: we (1) infer embeddings for all documents of the Pile, (2) cluster… (excerpt from the dataset card on Hugging Face)
Dataset facts
- Hub id
- JeanKaddour/minipile
- Likes
- 155
- Downloads (last 30 days)
- 2,685
- Tasks
- Text generation, Fill mask
- Modality
- text
- Size
- 1M to 10M examples
- License
- Other license
- Access
- Open download
- Author
- JeanKaddour · Personal account
- Languages
- en
- Created
- Apr 9, 2023
- Last modified
- Jun 20, 2023