metadata
dataset_info:
features:
- name: id
dtype: int64
- name: par_logits
sequence: float16
splits:
- name: train
num_bytes: 1667312196
num_examples: 4791127
download_size: 1639493168
dataset_size: 1667312196
configs:
- config_name: default
data_files:
- split: train
path: data/train-*