metadata
dataset_info:
features:
- name: INSTRUCTION
dtype: string
- name: RESPONSE
dtype: string
splits:
- name: train
num_bytes: 490292138
num_examples: 3114000
download_size: 287449359
dataset_size: 490292138
configs:
- config_name: default
data_files:
- split: train
path: data/train-*