kgcnn.losses package

Submodules

kgcnn.losses.losses module

class kgcnn.losses.losses.BinaryCrossentropyNoNaN(*args, **kwargs)

Bases: keras.src.losses.losses.BinaryCrossentropy

call(y_true, y_pred)

class kgcnn.losses.losses.ForceMeanAbsoluteError(reduction='sum_over_batch_size', name='force_mean_absolute_error', squeeze_states: bool = True, find_padded_atoms: bool = True, dtype=None)

Bases: keras.src.losses.loss.Loss

call(y_true, y_pred)

get_config()

class kgcnn.losses.losses.MeanAbsoluteError(reduction='sum_over_batch_size', name='mean_absolute_error', dtype=None)

Bases: keras.src.losses.loss.Loss

call(y_true, y_pred)

get_config()

Module contents