metadata
dataset_info:
features:
- name: json
struct:
- name: episode_id
dtype: string
- name: function_name
dtype: string
- name: height
dtype: int64
- name: messages
dtype: string
- name: parent
dtype: string
- name: split
dtype: string
- name: split_info
dtype: string
- name: step_id
dtype: int64
- name: uid
dtype: string
- name: width
dtype: int64
- name: png
dtype: image
- name: __key__
dtype: string
- name: __url__
dtype: string
splits:
- name: train
num_bytes: 10413293546.93212
num_examples: 14254
- name: test
num_bytes: 495759276
num_examples: 729
download_size: 10681551317
dataset_size: 10909052822.93212
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*