metadata
dataset_info:
features:
- name: rem
dtype: string
- name: add
dtype: string
- name: context
dtype: string
- name: meta
dtype: string
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 14814811902
num_examples: 2728295
- name: test
num_bytes: 3704062611
num_examples: 681983
download_size: 5172322839
dataset_size: 18518874513
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*