metadata
dataset_info:
features:
- name: id
dtype: string
- name: comment_text
dtype: string
- name: toxic
dtype: int64
- name: severe_toxic
dtype: int64
- name: obscene
dtype: int64
- name: threat
dtype: int64
- name: insult
dtype: int64
- name: identity_hate
dtype: int64
- name: input_word_ids
dtype: string
- name: input_mask
dtype: string
- name: all_segment_id
dtype: string
splits:
- name: train
num_bytes: 415834729
num_examples: 223549
download_size: 114181160
dataset_size: 415834729
configs:
- config_name: default
data_files:
- split: train
path: data/train-*