metadata
dataset_info:
features:
- name: image
dtype: image
- name: question
dtype: string
- name: ground
dtype: string
- name: prediction
dtype: string
- name: similarity
dtype: float64
splits:
- name: train
num_bytes: 3820497
num_examples: 100
download_size: 3775544
dataset_size: 3820497
configs:
- config_name: default
data_files:
- split: train
path: data/train-*