---
dataset_info:
  features:
  - name: pixel_values
    sequence:
      sequence:
        sequence: float32
  - name: input_ids
    sequence: int32
  - name: attention_mask
    sequence: int8
  - name: labels
    dtype: int64
  splits:
  - name: train
    num_bytes: 8147050794
    num_examples: 5994
  - name: test
    num_bytes: 7875210594
    num_examples: 5794
  download_size: 4581320669
  dataset_size: 16022261388
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---