README.md

metadata

dataset_info:
  features:
    - name: content
      dtype: string
  splits:
    - name: train
      num_bytes: 48490123196
      num_examples: 22690306
  download_size: 5070161762
  dataset_size: 48490123196
license: apache-2.0
task_categories:
  - text-generation
language:
  - en
size_categories:
  - 10M<n<100M

Dataset Card for "pretrain_en"

Tigerbot pretrain数据的英文部分。

Usage

import datasets

ds_sft = datasets.load_dataset('TigerResearch/pretrain_en')