metadata
dataset_info:
features:
- name: inputs
sequence:
sequence:
sequence: float32
- name: outputs
sequence: float32
splits:
- name: train
num_bytes: 207888000
num_examples: 2000
download_size: 76808567
dataset_size: 207888000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*