metadata
dataset_info:
features:
- name: input
dtype: string
- name: instruction
dtype: string
- name: idx
dtype: string
- name: output
dtype: string
- name: doc
dtype: string
splits:
- name: train
num_bytes: 3467473
num_examples: 4435
download_size: 975014
dataset_size: 3467473
configs:
- config_name: default
data_files:
- split: train
path: data/train-*