Hi,

I would like to double check the following.

The sample submission file is of the format

Image_ID | class | confidence | ymin | xmin | ymax | xmax

that is, one row represents the prediction and coordinates of one fruit in the image. For images with multiple fruits, should we have multiple rows (duplicate the image ID ) for each of the fruit prediction as it is the case in the Train.csv file? Are you able to share how your evaluation function will handle such cases?