-
Notifications
You must be signed in to change notification settings - Fork 1.1k
New_Home
Welcome to the MONAI (Medical Open Network for AI) wiki!
This wiki is intended to fulfill a number of purposes for both users of and collaborators on the MONAI codebase. Whether you are looking at using existing MONAI functionality, writing your own specialised functionality in a MONAI style, or even looking to contribute new functionality to MONAI, there are sections of the wiki to help you do just that.
The wiki is organised as follows:
- Developer guides
- Design
- Design philosophy
- Design discussions
- Contribution guidelines
- Continuous integration and delivery pipeline
Developer Guides give you detailed explanations and examples of MONAI features that can be used to make best use of MONAI, whether you are borrowing targeted parts of its functionality or building a MONAI application from scratch. MONAI is not an opinionated framework, and there is no "One True" MONAI way, but the developer guides provide you with examples that show you some of the effective ways of using the MONAI feature set.
Don't see what you need? Raise an issue so that we can continue adding to the developer guides. This wiki and MONAI rely on community feedback to be relevant and useful, so we are always happy to hear from you.
These are mainly provided in the form of example notebooks, that show MONAI usage in end to end tasks.
The following topics are (or will be) deep dives into how to use various aspects of the MONAI feature set
- Preprocessors and transforms
- Datasets and DataLoaders (Coming Soon)
- Networks (Coming Soon)
- Network layers (Coming Soon)
- Loss functions (Coming Soon)
- Metrics (Coming Soon)
- Engines and Loops (Coming Soon)
- Persisting and Restoring (Coming Soon)
MONAI is designed around the principle that the framework should be unopinionated and intuitive to use for both new users and those who are already familiar with pytorch and its related modules. As MONAI is an open source project and we are keen to encourage collaboration. As such the wiki has comprehensive documentation on the important design aspects of MONAI to help you get up to speed with the codebase and write new functionality that fits within the design and architecture of MONAI.
The main MONAI design page goes into more detail, but if you are interested in specific topics, the pages for those topics are listed below.
The following topics are (or will be) deep dives into the key design decisions for various aspects of MONAI:
- Preprocessors and transforms
- Datasets and DataLoaders (Coming Soon)
- Networks
- Network layers
- Loss functions (Coming Soon)
- Metrics (Coming Soon)
- Engines and Loops (Coming Soon)
- Persisting and Restoring (Coming Soon)
We believe that MONAI's design philosophy represents the right way to build a library for medicine-related deep learning. If you are interested in contributing to MONAI, we ask that you familiarize yourself with the MONAI philosophy, so that your contribution has the best chance of helping make MONAI successful.
The main MONAI design philosophy page covers this topic.
MONAI is still under active development, and design decisions are still being actively made. If you are interested in contributing to MONAI at the design level, you should be as familiar as possible with the ongoing design discussions as possible. Note, these pages are a little less curated than other pages on the wiki.
The following design discussion topic have their own pages:
- Preprocessors and transforms
- Datasets and DataLoaders (Coming Soon)
- Networks
- Network layers
- Loss functions (Coming Soon)
- Metrics
- Engines and Loops (Coming Soon)
- Persisting and Restoring (Coming Soon)
MONAI is an open source project and relies on contributions from our community to grow and thrive. The Contribution guidelines page covers everything that you need to know to contribute effectively to MONAI. We welcome contributions from the community; we only ask that you are familiar with the process.