Skip to content

Welcome to nnbench, a framework for reproducibly benchmarking machine learning models. The main goals of this project are portable and customizable benchmarking for ML models, and easy integration into existing ML pipelines.

Highlights:

  • Easy definition, bookkeeping and organization of machine learning benchmarks,
  • Enriching benchmark results with context to properly track and annotate results,
  • Streaming results to a variety of data sinks.