If the loaded datasets have varying epoch steps (e.g. ds1 show every 5th epoch an ds2 every 1st epoch) then the line/heatmap visualizations are not correct.
Before we implement this features, we have to check if we need it since we currently just have datasets with step size of 5 epochs.
We agreed on having datasets with equal epoch size and with equal granularity for now.
If this restriction is too strong then we can change the implementation accordingly in the future.
If the loaded datasets have varying epoch steps (e.g. ds1 show every 5th epoch an ds2 every 1st epoch) then the line/heatmap visualizations are not correct. Before we implement this features, we have to check if we need it since we currently just have datasets with step size of 5 epochs.