metadata
dataset_info:
features:
- name: image
dtype: image
- name: color
dtype:
class_label:
names:
'0': black
'1': blue
'2': cyan
'3': green
'4': purple
'5': red
'6': white
'7': yellow
- name: shape
dtype:
class_label:
names:
'0': circle
'1': diamond
'2': rectangle
'3': triangle
splits:
- name: train
num_bytes: 95525474
num_examples: 640
download_size: 95526422
dataset_size: 95525474
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: mit
task_categories:
- image-classification
language:
- en
tags:
- tutorial
size_categories:
- 10M<n<100M
sources:
- url: https://github.com/sdtemple/zootopia3
name: Simulation Script Source
This dataset is generated as a part of a tutorial on fitting PyTorch models. The tutorial is at github.com/sdtemple/zootopia3.
You can generate your own data with answers/simulation-exercise.ipynb at the repo.
Treat this as held-out test data, with 20 examples for each shape (4) and color (8) combo.