metadata
dataset_info:
features:
- name: task
dtype: string
- name: inputs
dtype: string
- name: targets
dtype: string
- name: symbols
sequence: string
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 185968803
num_examples: 70273
- name: validation
num_bytes: 212222
num_examples: 73
download_size: 88359630
dataset_size: 186181025
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*