nuprl/rkt-selfinstruct

--- dataset_info: features: - name: content dtype: string splits: - name: train num_bytes: 30686869 num_examples: 50048 download_size: 11737396 dataset_size: 30686869 configs: - config_name: default data_files: - split: train path: data/train-* license: openrail --- This is a self-instruction dataset for Racket code completion, with data generated using StarCoderBase-15B. The self-instruction recipe that we followed is the one here: https://huggingface.co/datasets/codeparrot/self-instruct-starcoder We have fine-tuned the following models with this dataset: - https://huggingface.co/nuprl/starcoderbase1b-rkt-selfinstruct

原始信息汇总

数据集概述

数据集信息

• 特征：
  • content：数据类型为字符串。

数据分割

• 训练集：
  • 名称：train
  • 大小：30,686,869字节
  • 示例数量：50,048

数据集大小

• 下载大小：11,737,396字节
• 数据集总大小：30,686,869字节

配置

• 配置名称：default
• 数据文件：
  • 分割：train
  • 路径：data/train-*

许可证

• 许可证类型：openrail