Skip to content

Latest commit

 

History

History
30 lines (23 loc) · 1.53 KB

README.md

File metadata and controls

30 lines (23 loc) · 1.53 KB

kauldron

Unittests PyPI version Documentation Status

Kauldron is a library for training machine learning models, optimized for research velocity and modularity.

Modularity:

  • All parts of Kauldron are self-contained, so can be used independently outside Kauldron.
  • Use any dataset (TFDS, Grain, SeqIO, your custom pipeline), any (flax) model, any optimizer,... Kauldron provides the glue that link everything together.
  • Everything can be customized and overwritten (e.g. sweep over models architecture, overwrite any inner layer parameter,...)

Research velocity:

  • Everything should work out-of the box. The example configs can be used and customized as a starting point.
  • Colab-first workflow for easy prototyping and fast iteration
  • Polished user experience (integrated XM plots, profiler, post-mortem debugging on borg, runtime shape checking, and many others...). Open an issue..

This is not an officially supported Google product.