---
dataset_info:
  features:
  - name: inputs
    sequence:
      sequence: float32
  - name: outputs
    sequence:
      sequence:
        sequence:
          sequence: float32
  splits:
  - name: train
    num_bytes: 7066533888
    num_examples: 2048
  download_size: 7012812883
  dataset_size: 7066533888
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---