metadata
dataset_info:
features:
- name: source
dtype: string
- name: tokens
sequence: int64
- name: index
dtype: int64
- name: text
dtype: string
- name: transcript
dtype: string
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 33934622038
num_examples: 2420047
download_size: 7263587230
dataset_size: 33934622038
configs:
- config_name: default
data_files:
- split: train
path: data/train-*