ertk.train.TrainConfig

class ertk.train.TrainConfig(balanced: bool = True, reps: int = 1, normalise: str = 'online', transform: TransformClass = TransformClass.std, seq_transform: str = 'global', n_jobs: int = 1, verbose: int = 0, label: str = 'label', sklearn: Any | None = None, pytorch: Any | None = None, tensorflow: Any | None = None)

Bases: ERTKConfig

Class to hold training configuration.

__init__(balanced: bool = True, reps: int = 1, normalise: str = 'online', transform: TransformClass = TransformClass.std, seq_transform: str = 'global', n_jobs: int = 1, verbose: int = 0, label: str = 'label', sklearn: Any | None = None, pytorch: Any | None = None, tensorflow: Any | None = None) → None

Methods

__init__([balanced, reps, normalise, ...])

Inherited Methods

from_config(config)	Create config object from any compatible config.
from_file(path[, override])	Create config from YAML file and optionlly override some values.
merge_with_args([args])	Merge config with command-line arguments.
to_dictconfig()	Convert config to DictConfig.
to_file(path)	Write config to YAML file.
to_string()	Generate YAML string representation of config.

Attributes

balanced	Whether to use class-balanced sample weights.
label	Name of label column.
n_jobs	Number of jobs to run in parallel.
normalise	Normalisation method.
pytorch	PyTorch configuration.
reps	Number of repetitions.
seq_transform	Transform method for sequence normalisation.
sklearn	Scikit-learn configuration.
tensorflow	TensorFlow configuration.
transform	Transform method for normalisation.
verbose	Verbosity level.

balanced: bool = True

Whether to use class-balanced sample weights.

label: str = 'label'

Name of label column.

n_jobs: int = 1

Number of jobs to run in parallel.

normalise: str = 'online'

Normalisation method. Can be one of “online”, or “none”.

pytorch: Any = None

PyTorch configuration.

reps: int = 1

Number of repetitions.

seq_transform: str = 'global'

Transform method for sequence normalisation.

sklearn: Any = None

Scikit-learn configuration.

tensorflow: Any = None

TensorFlow configuration.

transform: TransformClass = 'std'

Transform method for normalisation. Can be one of “std”, “minmax”, or “none”.

verbose: int = 0

Verbosity level.