metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 299233168
num_examples: 44876
- name: test
num_bytes: 65039672
num_examples: 9754
- name: dev
num_bytes: 13202640
num_examples: 1980
download_size: 116710696
dataset_size: 377475480
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: dev
path: data/dev-*