metadata
dataset_info:
features:
- name: label
dtype: int64
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 8783495
num_examples: 10000
- name: test
num_bytes: 8871330
num_examples: 10000
- name: dev
num_bytes: 4443760
num_examples: 5000
download_size: 8867398
dataset_size: 22098585
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: dev
path: data/dev-*