metadata
dataset_info:
features:
- name: inputs
dtype: large_string
- name: targets
dtype: large_string
- name: task
dtype: large_string
- name: index
dtype: int64
- name: inputs_nor
dtype: large_string
- name: inputs_backtranslation
dtype: large_string
- name: targets_nor
dtype: large_string
- name: targets_backtranslation
dtype: large_string
splits:
- name: train
num_bytes: 9331534698
num_examples: 2771562
- name: validation
num_bytes: 85141364
num_examples: 23860
- name: test
num_bytes: 2737659799
num_examples: 734178
download_size: 5880786502
dataset_size: 12154335861
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*