Mean Average Detection ignores warn_on_many_detections set to False

[shatalinra]

🐛 Bug

MeanAverageDetection still outputs warning about too many detections even if warn_on_many_detections set to False.

To Reproduce

The existing unit test could be used for reproduction, just try to set field to False before running it:
https://github.com/Lightning-AI/torchmetrics/blob/master/tests/unittests/detection/test_map.py#L715
The test will detect that warning is still outputted and therefore test passes.

Expected behavior

Warnings are not outputted at all.

Environment

Conda environment with Python 3.10 and torchmetrics 1.3.1 installed into it using pip.
See additional context why further details about environment are not actually important.

Additional context

I looked through sources and already identified the issue: https://github.com/Lightning-AI/torchmetrics/blob/master/src/torchmetrics/detection/mean_ap.py#L803
Basically, warn parameter is never used inside function. Instead it should be a condition for triggering warning.
Also another unit test case for the flag being set to False could be useful.

[github-actions]

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