How to get the iou of each class?

[zsc1104]

If i have 5 classes , i want to get the iou of each class , like the following code to get the acc of each class , how to use the torchmetrics.IoU ?
torchmetrics.Accuracy(num_classes=num_classes, average='none')

[SkafteNicki]

Hi, by setting the reduction parameter you can achieve this:

metric=IoU(num_classes=5, reduction='none')
target = torch.randint(0, 5, (10, 25, 25))
pred = torch.randint(0, 5, (10, 25, 25))
metric(pred, target)
tensor([0.1193, 0.1064, 0.1154, 0.1061, 0.1216])

[bw4sz]

@SkafteNicki from which module are you importing IoU, is this still valid?

>>> from torchmetrics.detection import IntersectionOverUnion
>>> metric  = IntersectionOverUnion(reduction="none")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/benweinstein/.conda/envs/DeepForest/lib/python3.9/site-packages/torchmetrics/detection/iou.py", line 151, in __init__
    super().__init__(**kwargs)
  File "/Users/benweinstein/.conda/envs/DeepForest/lib/python3.9/site-packages/torchmetrics/metric.py", line 146, in __init__
    raise ValueError(f"Unexpected keyword arguments: {', '.join(kwargs_)}")
ValueError: Unexpected keyword arguments: `reduction`

please see

#2152. I'd like to extract recall and precision for a given IoU threshold.