metadata
dataset_info:
features:
- name: messages
list: json
- name: tools
list: json
- name: source
dtype: string
- name: subset
dtype: string
- name: source_id
dtype: string
splits:
- name: train
num_bytes: 13319384573
num_examples: 500000
download_size: 13290001094
dataset_size: 13319384573
configs:
- config_name: default
data_files:
- split: train
path: data/train-*