Evaluation

In flash-ansr 0.6, evaluation, comparison baselines, and systematic benchmarking moved out of this repository into the standalone companion package srbf (Symbolic Regression Benchmark Framework).

Get srbf

pip install srbf

Repository: https://github.com/psaegert/srbf. See srbf's README for usage (running evaluations, configuring baselines, and external-model setup).

Evaluation configs

flash-ansr no longer ships evaluation configs. The configs/ tree now holds only the v23.* model bundles (plus test/ and test_set/ assets); the evaluation-config tree was removed. Evaluation configs and benchmark assets live in the srbf package, which drives them via FlashANSR.infer().

What stays in flash-ansr

• Model loading plus fit / predict via the FlashANSR API, see Getting Started.
• The flash_ansr benchmark CLI for dataset iteration timing.