metadata
dataset_info:
features:
- name: sample_id
dtype: int64
- name: category
dtype: string
- name: edit_instruction
dtype: string
- name: input_image
dtype: image
- name: target_image
dtype: image
- name: output_image_NanoBanana
dtype: image
- name: output_image_UniWorld
dtype: image
- name: output_checkpoint-3000_30_3.0
dtype: image
- name: output_checkpoint-3000_30_4.0
dtype: image
- name: output_image_ours_30_5.0
dtype: image
- name: output_checkpoint-3000_30_6.0
dtype: image
- name: caption
dtype: string
- name: qa_list
list:
- name: ground_truth_answer
dtype: string
- name: question
dtype: string
- name: new_column
dtype: string
- name: ours_30_5.0_list
list:
- name: Correction
dtype: bool
- name: answer
dtype: string
- name: ground_truth_answer
dtype: string
- name: question
dtype: string
- name: NanoBanana_list
list:
- name: Correction
dtype: bool
- name: answer
dtype: string
- name: ground_truth_answer
dtype: string
- name: question
dtype: string
- name: UniWorld_list
list:
- name: Correction
dtype: bool
- name: answer
dtype: string
- name: ground_truth_answer
dtype: string
- name: question
dtype: string
splits:
- name: train
num_bytes: 371211024
num_examples: 1001
download_size: 328402600
dataset_size: 371211024
configs:
- config_name: default
data_files:
- split: train
path: data/train-*