src.core.gray.grayWrapper module

Fit function Wrapper

src.core.gray.grayWrapper.flag_windows(wdw: Tensor, pdr_threshold: float, bdr_threshold: float, avg_ipt_threshold: float, std_ipt_threshold: float, skw_ipt_threshold: float, kur_ipt_threshold: float, log: LogHandler | Callable | None = None) → Tensor

src.core.gray.grayWrapper.get_emb(model: Model, data: Tensor, n_samples: int | None = None) → Tensor

get_emb.

Get the embeddings representation of an input.

Parameters:

model (Model) – The model to use for the embeddings.
data (tf.Tensor) – The data to embed.
n_samples (Optional[int]) – The number of samples to take. if None all the samples are taken.

src.core.gray.grayWrapper.gray_update(model: Model, pred_gray_triplets: DatasetV2, gray_targets: DatasetV2, training_data: Tuple[DatasetV2, DatasetV2, DatasetV2], training_triplets: Tuple[DatasetV2, DatasetV2, DatasetV2], gray_samples: Tensor, GrayTripletsGenerator: Callable, pdr_threshold: float | str | None = None, bdr_threshold: float | str | None = None, avg_ipt_threshold: float | str | None = None, std_ipt_threshold: float | str | None = None, skw_ipt_threshold: float | str | None = None, kur_ipt_threshold: float | str | None = None, gray_stats_output: str | None = None, batch_size: int | str = 0, logger: LogHandler = None) → Generator

src.core.gray.grayWrapper.gray_update_margin(model: Model, pred_gray_triplets: DatasetV2, training_data: Tuple[DatasetV2, DatasetV2, DatasetV2, DatasetV2], training_triplets: Tuple[DatasetV2, DatasetV2, DatasetV2, DatasetV2], gray_samples: Tensor, gray_exp_l: Tensor, GrayTripletsGenerator: Callable, margin: float | str = 0.0, gray_stats_output: str | None = None, gray_prediction_output: str | None = None, batch_size: int | str = 0, verbose: int = 1, logger: LogHandler | None = None) → Generator

src.core.gray.grayWrapper.gray_update_margin_freeze(model: Model, pred_gray_triplets: DatasetV2, gray_targets: DatasetV2, training_data: Tuple[DatasetV2, DatasetV2, DatasetV2], training_triplets: Tuple[DatasetV2, DatasetV2, DatasetV2], gray_samples: Tensor, GrayTripletsGenerator: Callable, margin: float | str = 0.0, sequence_threshold: int | str = 5, pdr_threshold: float | str | None = None, bdr_threshold: float | str | None = None, avg_ipt_threshold: float | str | None = None, std_ipt_threshold: float | str | None = None, skw_ipt_threshold: float | str | None = None, kur_ipt_threshold: float | str | None = None, gray_stats_output: str | None = None, gray_prediction_output: str | None = None, batch_size: int | str = 0, logger: LogHandler = None) → Generator

src.core.gray.grayWrapper.gray_update_margin_general(model, pred_gray_triplets: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, good: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, bad: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, training_triplets: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, gray: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, GrayTripletsGenerator: ~typing.Callable, *args, margin: float | str = 0.0, fixedMarginFlag: bool | str = False, gray_stats_output: str | None = None, gray_prediction_output: str | None = None, batch_size: int | str = 0, verbose: int = 0, logger=None, write_msg=<function f_logger.<locals>.__dummy_log>, **kwargs) → Generator

src.core.gray.grayWrapper.gray_update_margin_general_cure(model, pred_gray_triplets: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, good: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, bad: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, training_triplets: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, gray: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, GrayTripletsGenerator: ~typing.Callable, *args, margin: float | str = 0.0, fixedMarginFlag: bool | str = False, gray_stats_output: str | None = None, gray_prediction_output: str | None = None, gray_representors_output: str | None = None, batch_size: int | str = 0, n_reps: int = 5, n_samples_centr: int = 1000, verbose: int = 0, compression_factor: float = 0.5, ph: ~SNN2.src.io.pickleHandler.PickleHandler = None, logger=None, write_msg=<function f_logger.<locals>.__dummy_log>, **kwargs) → Generator

src.core.gray.grayWrapper.log_none(obj: str, msg: str, level: int = 0)

src.core.gray.grayWrapper.merge_triplets(new_data: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, old_data: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, *args, batch_size: int | None = None, logger=None, write_msg=<function f_logger.<locals>.__dummy_log>, **kwargs) → DatasetV2

src.core.gray.grayWrapper.predict(model: Model, dataset, **kwargs) → Tuple[ndarray, ndarray]

src.core.gray.grayWrapper.prediction_flag(distances: Tuple[ndarray, ndarray], margin: float = 0.0) → Tuple[Tensor, Tuple[ndarray, ndarray]]

src.core.gray.grayWrapper.regenerate_gray_triplets(predicted_flags: ~tensorflow.python.framework.tensor.Tensor, gray_samples: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, tripletGenerator: ~typing.Callable, goodSamples: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, badSamples: ~SNN2.src.decorators.decorators.c_logger.<locals>.augmented_cls, *args, portion: float = 1.0, logger=None, write_msg=<function f_logger.<locals>.__dummy_log>, **kwargs) → augmented_cls

src.core.gray.grayWrapper.remove_undecided(samples: Tensor, flags: Tensor) → Tuple[Tensor, Tensor]

src.core.gray.grayWrapper.save_gray_predictions(predictions: Tensor, distances: Tuple[ndarray, ndarray], output: str) → None

src.core.gray.grayWrapper.save_gray_stats(correct: Tensor, wrong: Tensor, output: str, undecided: bool = False, frozen_predictions: int | None = None) → None

src.core.gray.grayWrapper.save_representors(points: Tensor, output: str, ph: PickleHandler, label: str = 'Representors') → None

src.core.gray.grayWrapper.separate_training_triplets(training_data_samples: DatasetV2, training_data_classes: DatasetV2, log: LogHandler | None = None) → Tuple[DatasetV2, DatasetV2]

src.core.gray.grayWrapper.subIdx(idx: Tensor, flag: Tensor, reverse: bool = False) → Tensor

src.core.gray.grayWrapper.tf_prediction_flag(ap: Tensor, an: Tensor, margin: Tensor) → Tensor

src.core.gray.grayWrapper.validation_flag(distances: Tuple[ndarray, ndarray], expected_flags: Tensor, margin: float = 0.0, undecided_reverse: bool = False) → Tensor