metadata
dataset_info:
features:
- name: image
dtype: string
- name: conversation
struct:
- name: content
list:
- name: text
list: string
- name: type
list: string
- name: role
list: string
splits:
- name: train
num_bytes: 4676
num_examples: 8
- name: test
num_bytes: 527
num_examples: 1
- name: val
num_bytes: 548
num_examples: 1
download_size: 11937
dataset_size: 5751
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: val
path: data/val-*