Closed RobertoLorusso closed 10 months ago
In train and evaluation loops, after the dice scores for a batch are computed, they are inserted in a list called "metrics". This is wrongly initialised with the length of dataset instead of data loader. This hampers the mean values.
Join the discussion on DagsHub!
dagshub[bot]
commented
10 months ago
In train and evaluation loops, after the dice scores for a batch are computed, they are inserted in a list called "metrics". This is wrongly initialised with the length of dataset instead of data loader. This hampers the mean values.