Yet another machine learning framework.
##Guide
- To install:
pip install pymlx
- Demo
##Q&As
- Why not using existing open source packages?
- We do leverage some existing libraries (e.g. xgboost and even scikit-learn)
- None of the existing libraries satisfies all of our criteria. Here are some examples:
- Featurization framework: many machine learning libraries focus on the learning part but ignore the featurization step. In fact, most of so-called machine learning code in companies deals with featurization and in many cases, the learning part contains just a few lines of code.
- Model diagnostics: is there an easy way to analyze the performance of a model and see where it goes wrong or whether the labels are correct?
- More to come
- What's the OSS license of this library
- BSD3