NeuralTSNE.TSNE.CostFunctions package

Submodules

NeuralTSNE.TSNE.CostFunctions.cost_functions module

class CostFunctions(name)[source]View on GitHub

Bases: object

Class containing cost functions for t-SNE.

static kl_divergence(Y: Tensor, P: Tensor, params: dict[str, Any]) → Tensor[source]View on GitHub

Calculates the Kullback-Leibler divergence.

Parameters:

  • Y (torch.Tensor) – Embedding tensor.

  • P (torch.Tensor) – Conditional probability matrix.

Returns:

Kullback-Leibler divergence.

Return type:

torch.Tensor

Note

Calculates the Kullback-Leibler divergence between the true conditional probability matrix P and the conditional probability matrix Q based on the current embedding Y.

Module contents