---
dataset_info:
  features:
  - name: input_ids
    sequence: int32
  - name: attention_mask
    sequence: int8
  - name: pixel_values
    sequence:
      sequence:
        sequence:
          sequence: float64
  splits:
  - name: train
    num_bytes: 122737772355
    num_examples: 2660
  download_size: 20100863583
  dataset_size: 122737772355
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---