Fashionpedia Dataset

Name: FashionPedia Dataset
Creator: Menglin Jia, Mengyun Shi
License: https://creativecommons.org/publicdomain/zero/1.0/

Estimated reading: 3 minutes 792 views

Visualization of the Fashionpedia Dataset in the Deep Lake UI

Fashionpedia Dataset

What is Fashionpedia Dataset?

The Fashionpedia dataset consists of 48,825 clothing imagery in daily-life and celebrity event fashion labeled with complete segmentation for apparel and fine-grained features for segmented classes. Fashionpedia also includes an ontology created by fashion experts that include 27 main apparel classes, 19 apparel segments, 294 fine-grained attributes, and their relationships.

Download Fashionpedia Dataset in Python

Instead of downloading the Fashionpedia dataset in Python, you can effortlessly load it in Python via our Deep Lake open-source with just one line of code.

Load Fashionpedia Dataset Training Subset in Python

				
					import deeplake
ds = deeplake.load('hub://activeloop/fashionpedia-train')

Load Fashionpedia Dataset Testing Subset in Python

				
					import deeplake
ds = deeplake.load('hub://activeloop/fashionpedia-test')

Fashionpedia Dataset Structure

Fashionpedia Data Fields

images: tensor containing images.

images_meta: tensor containing meta data of images.

masks: tensor containing the masks.

boxes: tensor containing the bounding box coordinates.

categories: tensor containing the numerical labels that represent the index of the class in the category list.

super_categories: tensor containing the numerical labels that represent the index of the class in the super category list.

areas: tensor containing areas.

iscrowds: tensor containing the value that specifies if the image is crowd annotated or not.

attributes: tensor containing the attributes.

Fashionpedia Data Splits

The Fashionpedia dataset training set is composed of 45,623 samples.

The Fashionpedia dataset test set was composed of 2044 samples.

How to use Fashionpedia Dataset with PyTorch and TensorFlow in Python

Train a model on Fashionpedia dataset with PyTorch in Python

Let’s use Deep Lake built-in PyTorch one-line dataloader to connect the data to the compute:

				
					dataloader = ds.pytorch(num_workers=0, batch_size=4, shuffle=False)

Train a model on Fashionpedia dataset with TensorFlow in Python

				
					dataloader = ds.tensorflow()

Fashionpedia Dataset Creation

Data Collection and Normalization Information

Because the Fashionpedia ontology encompasses a wide range of fine-grained features for both clothes and clothing pieces, high-resolution pictures were picked for use in the curating process, which also helped in more accurate and quicker annotations for both segmentation masks and attributes. The Fashionpedia training pictures have an approximate width of 1710 and a height of 2151. While “masks” refer to a clothing instance that may have several independent components, “polygon” refers to a discrete region.

Additional Information about Fashionpedia Dataset

Fashionpedia Dataset Description

Homepage: https://fashionpedia.github.io/home/

Paper: https://arxiv.org/pdf/2004.12276.pdf

Point of Contact: Mengyun (David) Shi ([email protected]), Menglin Jia ([email protected]).

Fashionpedia Dataset Curators

Menglin Jia, Mengyun Shi, Mikhail Sirotenko, Yin Cui Google, Claire Cardie, Hartwig Adam, Bharath Hariharan, Van Dyk Lewis, Serge Belongie

Fashionpedia Dataset Licensing Information

Deep Lake users may have access to a variety of publicly available datasets. We do not host or distribute these datasets, vouch for their quality or fairness, or claim that you have a license to use the datasets. It is your responsibility to determine whether you have permission to use the datasets under their license. If you’re a dataset owner and do not want your dataset to be included in this library, please get in touch through a GitHub issue. Thank you for your contribution to the ML community!

Fashionpedia Dataset Citation Information

				
					@inproceedings{jia2020fashionpedia,
title={Fashionpedia: Ontology, Segmentation, and an Attribute Localization Dataset},
author={Jia, Menglin and Shi, Mengyun and Sirotenko, Mikhail and Cui, Yin and Cardie, Claire and Hariharan, Bharath and Adam, Hartwig and Belongie, Serge}
booktitle={European Conference on Computer Vision (ECCV)},
year={2020}
}