espnet2.spk.loss.softmax.Softmax

source

class espnet2.spk.loss.softmax.Softmax(nout: int, nclasses: int)

Bases: AbsLoss

Softmax loss

• Parameters:
  • nout – Dimension of input features (embedding size)
  • nclasses – Number of output classes

Initializes internal Module state, shared by both nn.Module and ScriptModule.

forward(input: Tensor, label: Tensor | None = None) → Tuple[Tensor | None, Tensor | None, Tensor]

Forward pass of Softmax loss.

• Parameters:
  • input – Input embeddings, shape (batch_size, embedding_dim)
  • label – Ground truth labels, shape (batch_size,)
• Returns: Cross-entropy loss accuracy: Classification accuracy preds: Predicted class indices
• Return type: loss