Override Defaults via Command-Line¶
The canari_ml train command uses a structured configuration system, where you can override default settings using CLI arguments or YAML config files.
Override Train Parameters¶
train.seed: Random seed for reproducibility.train.epochs: Number of training epochs.train.workers: Number of CPU/GPU workers for data loading and training.train.batch_size: Batch size for training.
Override Model Parameters¶
You can also override model-specific parameters:
model.network.filter_size: Filter size in the UNet architecture.model.litmodule.criterion.learning_rate: Learning rate for the optimiser.
Override Callbacks¶
callbacks.early_stopping.patience: Number of epochs to wait before applying early stopping.callbacks.model_checkpoint.monitor: Metric to monitor for saving checkpoints.
Examples¶
Basic Override Example¶
canari_ml train train.dataset=preprocessed_data/train_demo_dataset/03_cache_demo_dataset/cached.DAY.north.json train.name=demo_train train.epochs=2
Advanced Override Example¶
canari_ml train train.dataset=preprocessed_data/train_demo_dataset/03_cache_demo_dataset/cached.DAY.north.json train.name=demo_train train.seed=42 train.epochs=20 train.workers=8 train.batch_size=16
Override Callbacks¶
To modify callbacks like early stopping and checkpoint monitoring, add the following overrides to the above command: