:param vocab_dim: Number of possible values for W (and Y, if using), **excluding** mask dimension :param w_dims: The first dimensions

Returns the number of samples in the dataset.

Computes the precision@k for the specified values of k

Define dictionary of pre-compiled ProbLog models

Returns the weights dictionary used during ProbLog inference to update the graph semiring.

Build Worlds Queries Matrices The Kandinsky Figure has two pairs of objects with the same shape, in one pair the objects have the same co

Get concept probabilities for mcdropout Args: ensemble_c_prb_1 (ndarray): ensemble probability for concept 1 ensemble_c_prb_2 (nd

Function which computes a mean entropy per class Args: probabilities (ndarray): probability vector true_classes (ndarray): grount

Function which computes the class mean variance Args: probabilities (ndarray): probability vector true_classes (ndarray): grountr

Computes query probability given the worlds probability P(w). Args: self: instance query: query worlds_pr

Function which computes the concept accuracy Args: c1_prob (ndarray): first concept probability c2_prob (ndarray): second concept

:param t: The current time step :param s: The timestep to jump to :param x_0: The x at timestep 0 :param x_t: The x a

Substract mean from each vector, then raises to third power and compares cos similarity Does not modify any tensors in place

Substract mean from each vector, then raises to third power and compares cos similarity Does not modify any tensors in place Only compare

Create a fake stash, containing just the model name. This is used in general continual, as it is useless to backup a lightweight MNIST-360 training.

Creates the specified folder if it does not exist. Args: path: the complete path of the folder to be created Returns: None:

Method which creates checkpoint if it does not exists and loads it afterwards Args: model (nn.Module): model args: command line a

Creates the dictionary where to save the model status. Args: model: the model args: the current arguments dataset: the da

Defines the herbrand base to encode ProbLog worlds

Compute entropy given a probability distribution.

Computes the ECE filtering by index Args: confs (ndarray): confidence preds (ndarray): predictions labels (ndarray): labe

Arguments: x: Tensor, shape ``[seq_len, batch_size, embedding_dim]``