metadata
dataset_info:
features:
- name: transcript
dtype: string
- name: text
dtype: string
- name: compress_tokens
dtype: string
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 49099019908
num_examples: 3707675
download_size: 11985711300
dataset_size: 49099019908
configs:
- config_name: default
data_files:
- split: train
path: data/train-*