metadata
dataset_info:
features:
- name: text
dtype: string
- name: label
dtype: int64
- name: probs
dtype: float64
- name: metadata
dtype: string
- name: input_ids
sequence: int32
- name: token_type_ids
sequence: int8
- name: attention_mask
sequence: int8
- name: length
dtype: int64
- name: is_edu
dtype: int64
splits:
- name: train
num_bytes: 53510259794.11165
num_examples: 6558290
download_size: 28805694375
dataset_size: 53510259794.11165
configs:
- config_name: default
data_files:
- split: train
path: data/train-*