frgfm/imagewoof

Imagewoof is a subset of ImageNet containing ten dog‑breed categories, designed to provide a challenging image‑classification benchmark. Created by Jeremy Howard, it is released under the Apache‑2.0 license. The dataset includes images and corresponding labels, with a defined train/validation split.

Updated 12/11/2022

hugging_face

Dataset Overview

Basic Information

Name: Imagewoof
Creator: Jeremy Howard
License: Apache-2.0
Language: English
Size: 1K<n<10K
Task Type: Image Classification

Content

Summary

Imagewoof is an image dataset comprising ten categories, all of which are different dog breeds. It is intended to provide a relatively challenging image‑classification task.

Supported Tasks

Image Classification: Train image‑classification models.

Structure

Data Instances

Each data point consists of an image and its associated classification label.

{
  "image": <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=320x320 at 0x19FA12186D8>,
  "label": "Beagle"
}

Data Fields

image: a PIL.Image.Image object containing the image data.
label: the expected classification label for the image.

Splits

	train	validation
imagewoof	9025	3929

Creation

Source Data

Imagewoof is a subset of ImageNet. For details on source data collection and standardization, refer to the ImageNet documentation.

License

The dataset follows the Apache License 2.0.

Citation

@software{Howard_Imagewoof_2019,
    title={Imagewoof: a subset of 10 classes from Imagenet that arent so easy to classify},
    author={Jeremy Howard},
    year={2019},
    month={March},
    publisher = {GitHub},
    url = {https://github.com/fastai/imagenette#imagewoof}
}