Metrics¶
Evaluate your model.
-
tinder.metrics.semantic_segmentation_iou(score: torch.Tensor, label: torch.Tensor) → torch.Tensor[source]¶ Calculate IOU per class.
IOU (Intersection Over Union) = true positive / (TP + FP + FN) IOU = # (predict == label == class_i) / (predict==class_i or label==class_i)
- Parameters
prediction – [N, C, H, W] tensor
label – [N, H, W] tensor
- Returns
[N, C] tensor
- Return type
(torch.Tensor)