Why Zero2Neuro?
Our goals in developing Zero2Neuro are to:
-
Provide a low barrier-to-entry for new users to access deep learning tools in building predictive models using their own data and experimental questions
-
Provide a range of common model types that can be easily configured for use
-
Codify a set of best (good?) practices for performing machine learning experiments
How to Use this Documentation
This documentation is written for users with a range of knowledge. Top level material provides general introduction to the different topics. Deeper material is often tagged with anticipated knowledge level:
- Tutorial: detailed introduction to key concepts
- Beginner: what a new user needs to know to get started
- Intermediate: useful additions as a user begins to gain some experience with using Zero2Neuro
- Advanced: the powerful tools and options
What Zero2Neuro is Not
While our goal is for Zero2Neuro to have practical use in research pipelines, we do not pretend that this is a solution to all problems. Some users will find that Zero2Neuro does everything they need. Other users will find that with experience, they will grow out of Zero2Neuro and “strike out” on their own by writing their own Python code (either by creating code that plugs in to Zero2Neuro, using only some Zero2Neuro components, or developing entirely new code bases). All options are fine and encouraged!
We do not anticipate Zero2Neuro to solve problems in the areas of:
-
Large Language Models (though we support data involving sequences of tokens!)
-
Reinforcement Learning