metadata
dataset_info:
features:
- name: sentence
dtype: large_string
- name: codes_list
sequence:
sequence:
sequence: int64
- name: text_tokens
sequence: int64
- name: input_ids
sequence: int32
- name: labels
sequence: int64
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 22897293699
num_examples: 2223073
download_size: 6790203230
dataset_size: 22897293699
configs:
- config_name: default
data_files:
- split: train
path: data/train-*