README.md

---
dataset_info:
  features:
  - name: query
    dtype: string
  - name: image
    dtype: image
  - name: annot
    dtype: string
  - name: reasoning
    dtype: string
  - name: cate
    dtype: string
  - name: task
    dtype: string
  - name: metadata
    dtype: string
  - name: mask
    dtype: image
  - name: masks
    list: image
  splits:
  - name: train
    num_bytes: 363925192
    num_examples: 2674
  download_size: 286643659
  dataset_size: 363925192
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---