HASYv2 Dataset

Estimated reading: 4 minutes

Visualization of the HASYv2 dataset in the Deep Lake UI

HASYv2 dataset

What is HASYv2 Dataset?

HASYv2 Dataset is a dataset of single symbols (similar to MNIST). There are 168233 instances of 369 classes in HASY. It also contains two challenges: a classification challenge with 10 pre-defined folds for 10-fold cross-validation.

More than 150,000 handwritten symbols are contained in HASY. Similar to MNIST, HASY (the previous iteration of the dataset) is of very low resolution. Unlike MNIST, the HASYv2 dataset contains 369 classes, including Arabic numerals and Latin characters. Lastly, while MNIST is in grayscale, the HASYv2 dataset is in black and white.

Download HASYv2 Dataset in Python

Instead of downloading the HASYv2 dataset in Python, you can effortlessly load it in Python via our Deep Lake open-source with just one line of code.

Load HASYv2 Dataset Training Subset in Python

				
					import deeplake
ds = deeplake.load("hub://activeloop/hasy-train")

Load HASYv2 Dataset Testing Subset in Python

				
					import deeplake
ds = deeplake.load("hub://activeloop/hasy-test"

HASYv2 Dataset Structure

HASYv2 Data Fields

image: a tensor containing 32 x 32 pixel images.
latex: a class label tensor to classify the images into 369 classes of symbols.

HASYv2 Data Splits

HASYv2 training split comprises 151241 images.
HASYv2 testing split comprises 16992 images.

How to use HASYv2 Dataset with PyTorch and TensorFlow in Python

Train a model on HASYv2 dataset with PyTorch in Python

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

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

Train a model on HASYv2 dataset with TensorFlow in Python

				
					dataloader = ds.tensorflow()

Additional Information about HASYv2 Dataset

Homepage: https://zenodo.org/record/259444#.YfBN5ltBxGM
Repository: https://github.com/MartinThoma/HASY
Paper: “The HASYv2 dataset”. Martin Thoma(PDF).
Point of Contact: [email protected]
Dataset Curators: Martin Thoma

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!

Citation Information

				
					title={The HASYv2 dataset},
 author={Martin Thoma}, 
journal={arXiv preprint arXiv:1701.08380}, 
year={2017} }

HASYv2 Dataset FAQs

What is the HASYv2 dataset for Python?

HASYv2 is a dataset of more than 150,000 handwritten symbols. The HASYv2 dataset contains 369 classes including Latin characters, Arabic numerals, 32 different types of arrows, fractal and calligraphic Latin characters, brackets, and more. The images are black and white of the size 32 px × 32 px.

What is the HASYv2 dataset used for?

Like the MNIST dataset, HASYv2 is often used as a benchmark dataset, or as a proof of concept for training and testing purposes. It is often used to perform classification in the field of machine learning.

How to download the HASYv2 dataset in Python?

Using the open-source package Activeloop Deep Lake you can load the HASYv2 dataset fast with one line of code in Python. See detailed instructions on how to load the HASYv2 dataset training subset or the HASYv2 dataset testing subset in Python.

How can I use HASYv2 dataset in PyTorch or TensorFlow?

Using the open-source package Activeloop Deep Lake you can stream the HASYv2 dataset while training a model in PyTorch or TensorFlow with one line of code. See detailed instructions on how to train a model on the HASYv2 dataset with PyTorch in Python or train a model on the HASYv2 dataset with TensorFlow in Python.