MeanAveragePrecision() in torchmetrics==1.0.0 seems to be not working properly

[nisyad-ms]

🐛 Bug

Consider the example below where we try to compute mAP for a single threshold=0.2. The expected answer is 1.0, but we receive result['map'] = 0.0. torchmetrics v0.11.x gives the correct answer.

To Reproduce:

import torch
from torchmetrics.detection import MeanAveragePrecision

predictions = [dict(boxes=torch.tensor([[0.5, 0.5, 1, 1]]),
                    scores=torch.tensor([1.0]),
                    labels=torch.tensor([0]))]

targets = [dict(boxes=torch.tensor([[0, 0, 1, 1]]),
                labels=torch.tensor([0]))]

metric = MeanAveragePrecision(iou_thresholds=[0.2])
metric.update(predictions, targets)
result = metric.compute()  
result

[github-actions]

Hi! thanks for your contribution!, great first issue!