lisbet.cli.commands.train

Train a model for keypoint classification and export the embedder.

Functions

configure_export_embedder_parser(parser)	Configure export_embedder command parser.
configure_train_model_parser(parser)	Configure train_model command parser.
parse_data_augmentation(aug_string)	Parse data augmentation string into list of DataAugmentationConfig objects.
train_model(kwargs)	Train a model for keypoint classification.

lisbet.cli.commands.train.parse_data_augmentation(aug_string)

Parse data augmentation string into list of DataAugmentationConfig objects.

Parameters:

aug_string (str or None) – Comma-separated augmentation specifications, each with optional parameters. Format: name:p=value:frac=value Example: “all_perm_id:p=0.5,blk_perm_id:p=0.3:frac=0.2”

Returns:

List of dictionaries with augmentation configs, or None if None/empty.

Return type:

list[dict] or None

Examples

>>> parse_data_augmentation("all_perm_id")
[{'name': 'all_perm_id', 'p': 1.0}]

>>> parse_data_augmentation("all_perm_id:p=0.5,blk_perm_id:frac=0.3")
[{'name': 'all_perm_id', 'p': 0.5}, {'name': 'blk_perm_id', 'p': 1.0, 'frac': 0.3}]

>>> parse_data_augmentation("rotation:p=0.5:max_angle=30")
[{'name': 'rotation', 'p': 0.5, 'max_angle': 30.0}]

lisbet.cli.commands.train.configure_train_model_parser(parser)

Configure train_model command parser.

Return type:

None

lisbet.cli.commands.train.configure_export_embedder_parser(parser)

Configure export_embedder command parser.

Return type:

None

lisbet.cli.commands.train.train_model(kwargs)

Train a model for keypoint classification.