---
dataset_info:
  features:
  - name: url
    dtype: string
  - name: caption_llava
    dtype: string
  - name: caption_llava_short
    dtype: string
  - name: caption
    dtype: string
  - name: tags_open_images
    dtype: string
  - name: tags_booru
    dtype: string
  - name: key
    dtype: int64
  - name: clip_similarity_vitb32
    dtype: float64
  - name: clip_similarity_vitl14
    dtype: float64
  - name: nsfw_score_opennsfw2
    dtype: float64
  - name: nsfw_score_gantman
    dtype: float64
  - name: watermark_score
    dtype: float64
  - name: aesthetic_score_laion_v2
    dtype: float64
  - name: num_faces
    dtype: int64
  - name: width
    dtype: int64
  - name: height
    dtype: int64
  - name: exif
    dtype: string
  - name: sha256
    dtype: string
  splits:
  - name: train
    num_bytes: 3869799879.2561765
    num_examples: 2913654
  - name: validation
    num_bytes: 203673398.2428712
    num_examples: 153350
  - name: test
    num_bytes: 203674726.40327707
    num_examples: 153351
  download_size: 2168809755
  dataset_size: 4277148003.9023247
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---