dataset_info:
features:
- name: text
dtype: string
- name: id
dtype: string
- name: dump
dtype: string
- name: url
dtype: string
- name: file_path
dtype: string
- name: language
dtype: string
- name: language_score
dtype: float64
- name: token_count
dtype: int64
- name: score
dtype: float64
- name: int_score
dtype: int64
splits:
- name: train
num_bytes: 4909825141.472376
num_examples: 967210
download_size: 2894233408
dataset_size: 4909825141.472376
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: odc-by
Here is a sample README for your dataset:
1 Billion Tokens Subset of HuggingFaceFW/fineweb-edu
Overview
This dataset is a randomly sampled subset containing 1 billion tokens extracted from the sample-10BT dataset of the HuggingFaceFW/fineweb-edu corpus. It is intended for use in natural language processing research and educational purposes. The dataset has been filtered and processed to ensure a diverse and representative sample of the original dataset's content.
Source
- Original Dataset: HuggingFaceFW/fineweb-edu
- Subset:
sample-10BT
Description
The dataset comprises 1 billion tokens selected from the sample-10BT subset of the HuggingFaceFW/fineweb-edu collection.
Key Features
- Number of Tokens: 1 billion
- Format: Text data in a tokenized format
- Language: Primarily English
- License: ODC-BY 1.0 License
Usage
The dataset can be used for training language models, text classification, and other NLP tasks. Users are free to use, modify, and share the dataset under the terms of the ODC-BY license, which requires appropriate attribution.
Attribution
When using this dataset, please include the following citation:
This dataset is a subset of the sample-10BT dataset from HuggingFaceFW/fineweb-edu. The dataset is published under the ODC-BY 1.0 License.
License
This dataset is made available under the Open Data Commons Attribution License (ODC-BY) 1.0. You are free to:
- Share: Copy, distribute, and use the data.
- Adapt: Modify, transform, and build upon the data.
Conditions:
- Attribution: You must attribute the data in the manner specified by the original creators, but not in any way that suggests they endorse you or your use of the data.