train_gen2

Gen2 training entry point for the nested ERA5 data schema.

This script bypasses credit_main_parser entirely. It reads the config directly and assumes the new conf["data"]["source"] structure produced by MultiSourceDataset / BaseDataset.

For the legacy flat schema (v1), use applications/train.py.

Attributes

logger

Functions

main_cli()

Module Contents

train_gen2.logger

train_gen2.main_cli()