open-sourcing numerblox

Carlo Lepelaars

open-sourcing numerblox

Crowdcent, logo, investing, NLP, machine learning, numerai, crowdsourcing

Happy to announce we are open-sourcing a library at CrowdCent
called numerblox, which we have been building and using internally since January. We built it mainly to simplify and scale our weekly inference pipelines, but are also using components of it for training Numerai models.

The vision behind numerblox is that while the Numerai community builds models that may greatly differ from each other, our inference pipelines often have a similar structure:

1. Download new inference data each week
2. Preprocess
3. Make predictions
4. Postprocess
5. Evaluate
6. Submit

Building good data pipelines is mostly a software engineering effort, which currently every Numerai participant has to basically build from scratch. numerblox is designed to simplify every step in this process.

Installation: pip install numerblox

Github repo: