metadata
dataset_info:
features:
- name: category
dtype: string
- name: edit_instruction
dtype: string
- name: tar_img
dtype: image
- name: caption
dtype: string
- name: qa_list
list:
- name: ground_truth_answer
dtype: string
- name: question
dtype: string
- name: output_image_hidream_i1full
dtype: image
- name: output_image_mp_omnigen2
dtype: image
- name: output_image_Seedream
dtype: image
- name: output_image_3000_30_5.0
dtype: image
- name: output_image_flux
dtype: image
- name: output_image_bagel_think
dtype: image
- name: output_image_flux_kontext
dtype: image
- name: output_image_ovis_u1
dtype: image
- name: output_image_qwenimg
dtype: image
- name: output_image_GPT
dtype: image
- name: output_image_bagel_nothink
dtype: image
- name: output_image_UniWorld_1024
dtype: image
- name: output_image_Nanobanana
dtype: image
splits:
- name: train
num_bytes: 1558494076
num_examples: 2000
download_size: 1427958061
dataset_size: 1558494076
configs:
- config_name: default
data_files:
- split: train
path: data/train-*