---
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
  splits:
  - name: train
    num_bytes: 112
    num_examples: 4
  download_size: 3513
  dataset_size: 112
---

        method: cte_cold_emb_qv 

        num_inference_steps: 30

        {metric} : {mean} 
{metric} : {mean} 
{metric} : {mean} 
{metric} : {mean} 
{metric} : {mean}