README.md

---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: filepath
    dtype: string
  - name: sentids
    list: int32
  - name: filename
    dtype: string
  - name: imgid
    dtype: int32
  - name: split
    dtype: string
  - name: sentences_tokens
    list:
      list: string
  - name: sentences_raw
    list: string
  - name: sentences_sentid
    list: int32
  - name: cocoid
    dtype: int32
  - name: th_sentences_raw
    sequence: string
  splits:
  - name: train
    num_bytes: 988476410.9130616
    num_examples: 6000
  - name: validation
    num_bytes: 164746068.48551026
    num_examples: 1000
  - name: test
    num_bytes: 164746068.48551026
    num_examples: 1000
  download_size: 1316758173
  dataset_size: 1317968547.8840823
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---