metadata
dataset_info:
features:
- name: post_id
dtype: string
- name: source_image
dtype: image
- name: instruction
dtype: string
- name: rewritten_instruction
dtype: string
- name: creativity_level
dtype:
class_label:
names:
'0': Low
'1': Medium
'2': High
- name: objects
sequence: string
- name: object_categories
sequence: string
- name: primary_category
dtype: string
splits:
- name: train
num_bytes: 1602381079.11
num_examples: 3115
download_size: 1562686803
dataset_size: 1602381079.11
configs:
- config_name: default
data_files:
- split: train
path: data/train-*