metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: answer
dtype: string
- name: input_ids
sequence: int32
- name: labels
sequence: int64
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 7190631
num_examples: 500
download_size: 737993
dataset_size: 7190631
configs:
- config_name: default
data_files:
- split: train
path: data/train-*