metadata
dataset_info:
features:
- name: images
list: image
- name: problem
dtype: string
- name: answer
dtype: string
- name: edit_inst
dtype: string
- name: src_caption
dtype: string
- name: dst_caption
dtype: string
- name: text_score
dtype: float64
- name: fidelity_score
dtype: float64
- name: quality_score
dtype: float64
- name: overall_score
dtype: float64
splits:
- name: train
num_bytes: 2355224292
num_examples: 2984
- name: test
num_bytes: 788492172
num_examples: 1000
download_size: 3141276128
dataset_size: 3143716464
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*