About EvalSuite
What it is
EvalSuite brings machine learning, clinical, statistical, segmentation and object-detection evaluation into one framework with a shared API, structured results and publication-ready reporting.
Why it exists
To make evaluation easier to do correctly: fewer silent convention mismatches, uncertainty reported by default, and every metric documented with its assumptions and limitations.
Who it is for
Machine learning and computer vision researchers, clinical AI researchers, statisticians, data scientists, students, educators and reviewers who need to check how numbers were produced.
Philosophy
Scientific correctness comes before speed. Numerical edge cases are surfaced, not hidden. Claims about performance or validity are made only with evidence.
Open source
The package is released under the MIT License, with public issue tracking and a changelog. Source code: github.com/mkcs28/evalsuite-python.
Credits
Authors and maintainers: Manoj Kumar C S and Nikhil D Bharadwaj.
Roadmap
Core metrics and infrastructure shipped in v0.1.0, clinical and statistical evaluation in v0.2.0, computer vision in v0.3.0 and LLM evaluation in v0.4.0. See the roadmap.