jcarbonnell/preTrainingNEAR

---
language:
- en
license: apache-2.0
dataset_info:
features:
- name: text
dtype: string
splits:
- name: train
num_bytes: 6855034
num_examples: 1022
- name: val
num_bytes: 850752
num_examples: 114
download_size: 3998284
dataset_size: 7705786
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*
---
This dataset is a subset of the original nearData dataset, prepared for the continued pre-training of a pre-trained LLM.
The idea behind the continued pre-training of pre-trained models is to further instruct them with specific information, in this case on the Near Protocol blockchain, before fine-tuning them.

The preTrainingNEAR dataset was prepared from local text files using the datasets library from Hugging Face. It includes:

- nearBlog: 481 blog articles from Near Blog collected on March 13th, 2024.
- nearBosWebEngine: 13 docs files from the Near BOS Wen Engine collected on May 21st, 2024.
- nearDocs: 395 docs files from Near Docs collected on March 13th, 2024.
- nearNEPs: 124 docs files from the NEAR Enhancement Protocol collected on May 21st, 2024.
- nearNode: 40 docs files from the Near Node Docs collected on May 21st, 2024.
- nearPapers: 3 papers from the Near Papers collected on May 21st, 2024.
- nearWiki: 98 docs from the Near Wiki collected on May 21st, 2024.

原始信息汇总

数据集概述

基本信息

• 语言: 英语
• 许可证: Apache-2.0

数据集特征

• 特征名称: text
• 数据类型: string

数据集划分

• 训练集
  • 样本数量: 1022
  • 数据大小: 6855034 字节
• 验证集
  • 样本数量: 114
  • 数据大小: 850752 字节

数据集大小

• 下载大小: 3998284 字节
• 总数据大小: 7705786 字节

数据文件配置

• 默认配置
  • 训练集路径: data/train-*
  • 验证集路径: data/val-*

数据来源

• nearBlog: 481篇博客文章
• nearBosWebEngine: 13篇文档
• nearDocs: 395篇文档
• nearNEPs: 124篇文档
• nearNode: 40篇文档
• nearPapers: 3篇论文
• nearWiki: 98篇文档