fl_sim.models.top_n_accuracy#
- fl_sim.models.top_n_accuracy(preds: Tensor, labels: Tensor, n: int = 1) float[source]#
Top-n accuracy.
- Parameters:
preds (torch.Tensor) – Shape
(batch_size, n_classes)or(batch_size, n_classes, d_1, ..., d_n).labels (torch.Tensor) – Shape
(batch_size,)or(batch_size, d_1, ..., d_n).
- Returns:
The top-n accuracy.
- Return type: