What is modelhub?

Welcome to modelhub!

Thank you for offering to share your deep learning experiences.

Modelhub is a collection of deep learning models for medical data and associated meta-analysis. It aims to:

• Establish an online open-access collection of successful deep learning models - covering a wide range of applications on medical data ranging anywhere from radiographs to genomics and pathology
• Extract meta-analysis from the model collection and offer insights into trends, styles and best practices - as a means to alleviate the redundant trial & error nature of deep learning research
• Promote reproducible research, transfer learning efforts, as well as increase the visibility and exposure of contributors' work.

To contribute a model, please fill in this form. The form should take under 30 minutes to fill in. In addition to providing a link to the model in these formats, the form also asks for general information helpful to anyone interested in downloading and fine-tuning the models to fit their own data - such as data preprocessing steps and hyperparameters used. While sharing training data and full code implementations of your efforts are not required, they are encouraged if plausible. Transfer learning approaches with fine-tuned models are also welcome.

All rights to the models and related information remain with their respective authors and can be taken offline at the authors' request. By default, all models will be made available under the MIT licence. However, feel free to specify any license you deem appropriate for your work.

We are very excited to introduce this resource to the community, which would not be possible without your valuable contribution. We believe this resource will be of great value to the community as a whole, promote reproducible research and give your work greater visibility.

About us: We are the Computational Imaging and Bioinformatics Laboratory at the Harvard Medical School, Brigham and Women’s Hospital and Dana-Farber Cancer Institute. We are a data science lab focused on the development and application of novel Artificial Intelligence (AI) approaches to various types of medical data.