metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 8709278
num_examples: 176
download_size: 954382
dataset_size: 8709278
configs:
- config_name: default
data_files:
- split: train
path: data/bison/train-*