metadata
language:
- en
multilinguality:
- monolingual
size_categories:
- 1M<n<10M
task_categories:
- feature-extraction
- sentence-similarity
pretty_name: Wikipedia Sentences
tags:
- sentence-transformers
dataset_info:
features:
- name: sentence
dtype: string
splits:
- name: train
num_bytes: 958187457
num_examples: 7871825
download_size: 683051610
dataset_size: 958187457
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
Dataset Card for Wikipedia Sentences (English)
This dataset contains 7.87 million English sentences and can be used in knowledge distillation of embedding models.
Dataset Details
- Columns: "sentence"
- Column types:
str
- Examples:
{ 'sentence': "After the deal was approved and NONG's stock rose to $13, Farris purchased 10,000 shares at the $2.50 price, sold 2,500 shares at the new price to reimburse the company, and gave the remaining 7,500 shares to Landreville at no cost to him." }
- Collection strategy: Downloaded from https://sbert.net/datasets/wikipedia-en-sentences.txt.gz and uploaded without further modifications
- Deduplified: No