metadata
dataset_info:
features:
- name: img_path
dtype: string
- name: expected_emotion
dtype: string
- name: prompt
dtype: string
- name: prediction
dtype: string
- name: explanation
dtype: string
- name: is_correct
dtype: int32
splits:
- name: train
num_bytes: 204959
num_examples: 140
download_size: 37838
dataset_size: 204959
configs:
- config_name: default
data_files:
- split: train
path: data/train-*