metadata
dataset_info:
features:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: state
sequence: string
- name: tokens
dtype: int64
splits:
- name: train
num_bytes: 2858565332.9004607
num_examples: 2880075
download_size: 854216354
dataset_size: 2858565332.9004607
configs:
- config_name: default
data_files:
- split: train
path: data/train-*