---
dataset_info:
  features:
  - name: index
    dtype: string
  - name: url
    dtype: string
  - name: raw_prompt
    dtype: string
  - name: ground_truth
    dtype: image
  - name: gen_images_sd1_4
    list: image
  - name: minimal_rephrased_prompt
    dtype: string
  - name: gen_images_minimal_rephrased
    list: image
  - name: mse_minimal_rephrased
    dtype: float32
  - name: gpt_rephrased_prompt
    dtype: string
  - name: gen_images_gpt_rephrased
    list: image
  - name: mse_gpt_rephrased
    dtype: float32
  - name: qwen_recaptioned_prompt
    dtype: string
  - name: gen_images_qwen_recaptioned
    list: image
  - name: mse_qwen_recaptioned
    dtype: float32
  splits:
  - name: train
    num_bytes: 3036242326.0
    num_examples: 500
  download_size: 3023380977
  dataset_size: 3036242326.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---