metadata
dataset_info:
features:
- name: content
dtype: string
- name: input_ids
sequence: int32
- name: ratio_char_token
dtype: float64
- name: token_count
dtype: int64
splits:
- name: train
num_bytes: 39808133413
num_examples: 12725978
download_size: 15807072713
dataset_size: 39808133413
configs:
- config_name: default
data_files:
- split: train
path: data/train-*