metadata
dataset_info:
features:
- name: input_ids
list: int32
- name: labels
list: int64
- name: role_ids
list: int64
splits:
- name: train
num_bytes: 335981672
num_examples: 50661
download_size: 46783528
dataset_size: 335981672
configs:
- config_name: default
data_files:
- split: train
path: data/train-*