metadata
dataset_info:
features:
- name: input_ids
sequence: int32
length: 405
- name: label
dtype:
class_label:
names:
'0': contact
'1': corners
'2': culture
'3': detect
'4': frame
'5': grow
'6': half_fill
'7': replace_color
'8': translate
splits:
- name: train
num_bytes: 1628227920
num_examples: 1000140
download_size: 104785279
dataset_size: 1628227920
configs:
- config_name: default
data_files:
- split: train
path: data/train-*