I would like to create a report not just with graph plots of my metrics but also a table with the best metric results (accuracy, mAP, …). So that the viewer does not have to compare each line of a run, but can instead look at the table with the scalar best results if he wants to identify the best run.

It seems to me as there is currently no such feature implemented?

We do provide such a feature, it is called a Scalar Chart. You should be able to add a scalar chart to any panel by clicking the top right of the panel and selecting “Scalar Chart”. You can then define the metric you want to consider in that one chart.

Sorry, but that is not really what I want. The scalar chart can only be used for tracked parameters and then aggregates over all values of that parameter for every run.
The result is always just a single value for all runs combined.
Example: I select the parameter accuracy and choose “Agg=Max”. Then the mean accuracy for every run is computed and then the maximum mean value of all runs is chosen.
This is not what I want. I want that the max of a single run is chosen and plotted in a table along the max values of every other run.

This is a workflow that we’ll be making much easier in the future, making it easier to get min, max etc. for a given metric for a given run / group of runs. In the meantime, you can just log the max value at the end of your runs and get the median of that. We have a convenience method called define_metric which you can use to automatically do this for you.

wandb.init(entity="wandb", project="define-metric-demo")
# define a metric we are interested in the minimum of
wandb.define_metric("loss", summary="min")
# define a metric we are interested in the maximum of
wandb.define_metric("acc", summary="max")

This will then log the max/min value of a given logged metric for you.
Here’s more documentation of this method: