metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
dtype: int64
splits:
- name: train
num_bytes: 1827789
num_examples: 7999
- name: eval
num_bytes: 527000
num_examples: 2000
download_size: 569069
dataset_size: 2354789
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: eval
path: data/eval-*