Unable to track multiple types of COCOMetric during training

[nicjac]

🐛 Bug

Describe the bug When multiple instances of COCOMetric are used for different metric_type, only one is tracked during training

To Reproduce Steps to reproduce the behavior:

1. Create a learner object with two different instances of COCOMetric

metrics=[COCOMetric(metric_type=COCOMetricType.mask), COCOMetric(metric_type=COCOMetricType.bbox)]

2. During training, only a single COCOMetric is tracked

Expected behavior The different instances of COCOMetric should be tracked independently (e.g. in cases where both mask and bounding box metrics are of interest)

[FacuRoffet99]

It works fine for me. This is the code I'm using:

metrics = [COCOMetric(metric_type=COCOMetricType.mask), COCOMetric(metric_type=COCOMetricType.bbox)]
learn = model_type.fastai.learner(dls=[train_dl, valid_dl], model=model, metrics=metrics)

Maybe you forgot to update the metrics in the learner itself (second line)?