JeanKaddour/minipile
Dataset Card for MiniPile Dataset Description 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 the embedding space using k-means, and (3) filter out low-quality clusters. The primary motivation for curating MiniPile is that… See the full description on the dataset page: https://huggingface.co/datasets/JeanKaddour/minipile.
| Author / organization | JeanKaddour |
|---|---|
| License (as declared) | other |
| Downloads | 2.8K (as reported by Hugging Face) |
| Likes | 155 |
| Created | 9 April 2023 |
| Last updated at source | 20 June 2023 |
Popularity, last 2 days
2,802 now (-92 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
- afrideva/smol_llama-220M-GQA-GGUFModel · used to train
- mradermacher/llama3-42b-v0-i1-GGUFModel · used to train
- mradermacher/HelpingAI2-9B-i1-GGUFModel · used to train
- Locutusque/TinyMistral-248MModel · used to train
- bartowski/HelpingAI2-9B-GGUFModel · used to train
- HelpingAI/HelpingAI2-6BModel · used to train
- BEE-spoke-data/smol_llama-101M-GQAModel · used to train
- mradermacher/Llama-3-5B-Sheard-GGUFModel · used to train
- afrideva/smol_llama-101M-GQA-GGUFModel · used to train
- mradermacher/HelpingAI2-9B-GGUFModel · used to train
- mradermacher/llama3-42b-v0-GGUFModel · used to train
- QuantFactory/HelpingAI2-9B-GGUFModel · used to train
- mradermacher/TinyMistral-248M-GGUFModel · used to train
- mradermacher/smol_llama-4x220M-MoE-i1-GGUFModel · used to train
- stabilityai/stablelm-base-alpha-3b-v2Model · used to train
- mradermacher/TinyMistral-248M-i1-GGUFModel · used to train
- BEE-spoke-data/mega-ar-126m-4kModel · used to train
- mradermacher/HelpingAI2-6B-i1-GGUFModel · used to train
- mradermacher/smol_llama-4x220M-MoE-GGUFModel · used to train
- HelpingAI/HelpingAI2-9BModel · used to train
- mirth/chonky_modernbert_large_1Model · used to train
- QuantFactory/TinyMistral-248M-v3-GGUFModel · used to train
- afrideva/TinyMistral-248M-GGUFModel · used to train
- mradermacher/HelpingAI2-6B-GGUFModel · used to train
- mirth/chonky_mmbert_small_multilingual_1Model · used to train
- QuantFactory/HelpingAI2-6B-GGUFModel · used to train
- stabilityai/stablelm-base-alpha-7b-v2Model · used to train
- BEE-spoke-data/smol_llama-220M-GQAModel · used to train
- mradermacher/TinyMistral-248M-v3-GGUFModel · used to train
Recent changes
- 16 Oct 2026 Description: Dataset Card for MiniPile Dataset Description 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 f → Dataset Card for MiniPile Dataset Description 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 f
Changes detected by Silytics between refreshes of the source.
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.