metadata
dataset_info:
features:
- name: id
dtype: string
- name: translation
dtype:
translation:
languages:
- vi
- zh
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 5673312936
num_examples: 13329734
- name: test
num_bytes: 630033778
num_examples: 1481082
download_size: 3190585179
dataset_size: 6303346714
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*