metadata
dataset_info:
features:
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
- name: language
dtype: string
- name: origin
dtype: string
- name: original_id
dtype: string
- name: instruction_type
dtype: string
- name: instruction_translated
dtype: bool
- name: output_type
dtype: string
- name: output_translated
dtype: bool
splits:
- name: train
num_bytes: 5116737439
num_examples: 1545169
download_size: 5019105158
dataset_size: 5116737439
configs:
- config_name: default
data_files:
- split: train
path: data/train-*