t_vbll_loss

probly.losses.torch.t_vbll_loss(layer: TVBLLLayer, features: Tensor, targets: Tensor, regularization_weight: float) → Tensor[source]

Negative Student-t VBLL ELBO from [HWS24].

Implements the Student-t discriminative objective of a TVBLLLayer, combining the reduced Knowles-Minka softmax bound with the Gamma noise-precision KL and the weight-posterior KL.

Parameters:

  • layer – The Student-t variational Bayesian last layer to fit.

  • features – Backbone features feeding the layer, shape (batch, in_features).

  • targets – Integer class labels, shape (batch,).

  • regularization_weight – Weight on the regularization terms (typically 1 / dataset_size).

Returns:

A scalar tensor with the negative ELBO to minimize.