metadata
dataset_info:
features:
- name: task
dtype: string
- name: env_seed
dtype: int64
- name: image
dtype: image
- name: instruction
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 30915145.25
num_examples: 2198
- name: validation
num_bytes: 863619
num_examples: 100
- name: test
num_bytes: 802233
num_examples: 100
download_size: 3181061
dataset_size: 32580997.25
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*