Embedded Files
machine learning engine
The cosmosis machine learning engine combines the most powerful open-source libraries for data science and machine learning while remaining light weight, modular and extendable.
This framework is designed to reduce boilerplate and provide a simple and flexible form and tools with which to construct data science pipelines.
The conda environment can be set up with just a few command lines.
A dataset can be created by simply implementing the load_data() method of the CDataset class.
In your environment's jupyter lab input the parameters and create the learner.
Pipelines can be created using default, off-the-self or custom components.
cosmosis is designed to be modular and light weight so that all of it's components may be easily inspected and modified.
Here is a link to the github experiments notebook showing a series of examples demonstrating how to use the framework.
Page updated
Google Sites
Report abuse