Ch/update mes example

[cleherny]

I've noticed a typo in the README of the MES example and an error occurred when running the GE swimming pool example.
For the latter, I get the following metrics:

2025-01-20 09:55:43 - INFO - Tagging detections with threshold = 0.60, which maximizes the f1-score on the val dataset.
2025-01-20 09:55:43 - INFO - Method to compute the metrics = micro-average
2025-01-20 09:55:55 - INFO - Dataset = trn => precision = 0.945, recall = 0.810, f1 = 0.872
2025-01-20 09:55:57 - INFO - Dataset = val => precision = 0.926, recall = 0.762, f1 = 0.836
2025-01-20 09:55:59 - INFO - Dataset = tst => precision = 0.910, recall = 0.782, f1 = 0.841
2025-01-20 09:56:03 - INFO - Dataset = oth => precision = 0.427, recall = 0.553, f1 = 0.482

Is this what we used to get (I don't remember)? In particular, the other metrics are much lower than the others.

[stdl-s-sonarqube]

Analysis Details

0 Issues

• 0 Bugs
• 0 Vulnerabilities
• 0 Code Smells

Coverage and Duplications

• No coverage information (0.00% Estimated after merge)
• 0.00% Duplicated Code (0.00% Estimated after merge)

Project ID: swiss-territorial-data-lab_object-detector_AYZ4zWIzr7JdaaSXwexc

View in SonarQube

[GwenaelleSa]

If you tested and are confident it works properly, it's fine by me.

[GwenaelleSa]

It should make no difference with one class.

[cleherny]

Yes it is the same.