Skip to main content

Datasets

Currently, to use fpm-py, you will need to work with your own data. We are actively developing public datasets that will be accessible directly within the library, similar to how Hugging Face's transformers library works.

This feature will allow you to experiment with the Fourier Ptychography algorithm without needing any hardware. These datasets and the easy-to-use download functionality will be available in a future release. Stay tuned!

Sample Dataset​

If you want to get started with fpm-py, you can use a sample ImageSeries dataset available on our GitHub repository.

You can find the sample ImageSeries dataset at the following link: Sample Dataset on GitHub.

To load the sample dataset, download it from GitHub and use the load() method provided by the ImageSeries class:

import fpm_py as fpm

# Load the sample dataset from GitHub
dataset = fpm.ImageSeries.load("path/to/example.pt")

# Reconstruct the high-resolution image
output = fpm.reconstruct(dataset)