metadata
dataset_info:
features:
- name: input_values
sequence: float32
- name: input_length
dtype: int64
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 22066238988
num_examples: 51583
download_size: 18951867322
dataset_size: 22066238988
configs:
- config_name: default
data_files:
- split: train
path: data/train-*