metadata
dataset_info:
features:
- name: original_sentence
dtype: string
- name: input_ids
sequence: int32
- name: target_id
dtype: int64
- name: attention_mask
sequence: int8
splits:
- name: train_chunk_4
num_bytes: 158116391
num_examples: 100000
download_size: 2544674
dataset_size: 158116391
configs:
- config_name: default
data_files:
- split: train_chunk_4
path: data/train_chunk_4-*