---
dataset_info:
  features:
  - name: key
    dtype: string
  - name: zenodo_record_id
    dtype: string
  - name: zenodo_filename
    dtype: string
  - name: page_number
    dtype: int64
  - name: text_embedding
    list: float32
  - name: visual_embedding
    list: float64
  - name: mixed_embedding
    list: float32
  - name: extracted_text
    dtype: string
  - name: structured_response
    dtype: string
  - name: mixed_embedding_structured
    list: float32
  splits:
  - name: train
    num_bytes: 44504350
    num_examples: 2617
  download_size: 33440765
  dataset_size: 44504350
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---