metadata
dataset_info:
features:
- name: title
dtype: string
- name: topic
dtype: string
- name: text
dtype: string
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
- name: packed_length
dtype: int64
splits:
- name: train
num_bytes: 95676495.29356267
num_examples: 26180
- name: validation
num_bytes: 5035989.706437333
num_examples: 1378
download_size: 13309382
dataset_size: 100712485
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*