Quickstart

The toy pipeline in configs/toy-pipeline.yaml runs end to end in about a minute on a laptop.

class-e generate configs/toy-pipeline.yaml        # train and test designs -> .npz
class-e status   data/toy/parts                   # chunks, points, failures
class-e train    configs/toy-pipeline.yaml        # -> bundles/toy.cemu, prints held-out accuracy
class-e evaluate bundles/toy.cemu --config configs/toy-pipeline.yaml \
                 --plots plots/toy --report plots/toy.html
class-e info     bundles/toy.cemu

Use the bundle from Python:

from class_e import EmulatorBundle

bundle = EmulatorBundle.load("bundles/toy.cemu")
tt = bundle.predict({"h": 0.68, "omega_b": 0.0224, "omega_cdm": 0.12})["TT"]

Tip

Add --dry-run to generate to print the settings without running CLASS, and --set KEY=VALUE to override any YAML value.

Next: the CLI reference and the YAML configuration.