---
dataset_info:
  features:
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 2058322608
    num_examples: 870460
  - name: validation
    num_bytes: 1033274
    num_examples: 438
  - name: test
    num_bytes: 20641182
    num_examples: 8651
  download_size: 1232517026
  dataset_size: 2079997064
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
license: mit
task_categories:
- text-generation
language:
- en
pretty_name: MiniPile-2048
size_categories:
- 100K<n<1M
---

# Minipile-2048

MiniPile-2048 is the pruned version of `JeanKaddour/minipile`. Prunining performed with Llama-3 tokenizer and everything longer than 2048 tokens are filtered out. 

This dataset is for the GPU poor. 

## Dataset Curators

MiniPile-2048 is a subset of MiniPile, curated by Berat Çimen. MiniPile is a subset of the Pile, curated by Jean Kaddour. The Pile was created by Leo Gao, Stella Biderman, Sid Black, Laurence Golding, Travis Hoppe, Charles Foster, Jason Phang, Horace He, Anish Thite, Noa Nabeshima, Shawn Presser, Connor Leahy.

## Licensing Information

Since MiniPile-2048 and MiniPile is a subset of the Pile, the same MIT License holds.