metadata
dataset_info:
features:
- name: json_input
dtype: 'null'
- name: titre
dtype: string
- name: prompt0
dtype: string
- name: prompt
dtype: string
- name: response
dtype: string
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: tokens
dtype: string
splits:
- name: train
num_bytes: 19649430
num_examples: 7566
download_size: 10235687
dataset_size: 19649430
configs:
- config_name: default
data_files:
- split: train
path: data/train-*