metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: label
dtype: string
- name: prompt_similarity
dtype: float32
- name: identity_consistency
dtype: float32
- name: negative_prompt_similarity
dtype: float32
- name: target_prompt_similarity
dtype: float32
- name: aesthetic_score
dtype: float32
- name: ir_score
dtype: float64
splits:
- name: train
num_bytes: 636
num_examples: 19
download_size: 4619
dataset_size: 636
method: tex_inv_cold_
num_inference_steps: 30
prompt_similarity : 0.24850334227085114
identity_consistency : 0.5155891180038452 negative_prompt_similarity : 0.1717078983783722 target_prompt_similarity : 0.19024232029914856 aesthetic_score : 3.796888589859009 ir_score : -1.1301129200809454