metadata
dataset_info:
features:
- name: id
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
- name: text
dtype: string
- name: is_toxic
dtype: int64
splits:
- name: train
num_bytes: 424890269
num_examples: 223549
download_size: 118843841
dataset_size: 424890269
configs:
- config_name: default
data_files:
- split: train
path: data/train-*