metadata
dataset_info:
features:
- name: task
dtype: string
- name: output
dtype: string
- name: instruction
dtype: string
splits:
- name: train
num_bytes: 10113244285.797682
num_examples: 553419
- name: validation
num_bytes: 18099538.178737093
num_examples: 1000
download_size: 5306619706
dataset_size: 10131343823.97642
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*