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: 69601717.00154574
num_examples: 78052
- name: validation
num_bytes: 3664139.9984542546
num_examples: 4109
download_size: 8705351
dataset_size: 73265857
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*