---
dataset_info:
  features:
  - name: id
    dtype: string
  - name: prompt
    dtype: string
  - name: latents
    dtype:
      array4_d:
        shape:
        - 2
        - 4
        - 64
        - 64
        dtype: float16
  - name: prompt_embed
    dtype:
      array3_d:
        shape:
        - 1
        - 77
        - 1024
        dtype: float16
  - name: aes_score
    dtype: float64
  splits:
  - name: train
    num_bytes: 13296434205.0
    num_examples: 58898
  download_size: 12803581610
  dataset_size: 13296434205.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---