greydanus/mnist1d

A 1D analogue of the MNIST dataset for measuring spatial biases and answering Science of Deep Learning questions.

55
/ 100
Established

This project provides a simplified 1D dataset, called MNIST-1D, for researchers and students working on deep learning models. It helps you quickly test and compare the performance of different neural network architectures, like CNNs versus MLPs, for classification tasks. The input is a 40-dimensional synthetic data sequence, and the output is a classification. It's designed for deep learning researchers, academics, and educators.

238 stars. Used by 1 other package. No commits in the last 6 months. Available on PyPI.

Use this if you need a lightweight, easily modifiable dataset to explore fundamental deep learning concepts, benchmark new model architectures, or teach about spatial inductive biases and model discrimination.

Not ideal if you need a complex, high-dimensional, or real-world dataset for production-ready applications or models that require fine-grained image recognition.

deep-learning-research neural-network-benchmarking educational-dataset model-comparison spatial-biases
Stale 6m
Maintenance 0 / 25
Adoption 11 / 25
Maturity 25 / 25
Community 19 / 25

How are scores calculated?

Stars

238

Forks

38

Language

Jupyter Notebook

License

Apache-2.0

Last pushed

Oct 09, 2024

Commits (30d)

0

Dependencies

4

Reverse dependents

1

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/greydanus/mnist1d"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.