Questions regarding Training(parameters)

[OrkunYilmaz]

Hello what kind of loss/error is Err1 and Err2 and the value in the brackets during training process?

Also could you please provide short informations about the following training parameters:

-backWeight -maxIter -retrain -nThreads -nEpochs -shareGradInput

Much thanks in advance

[XingangPan]

@OrkunYilmaz Err1 is semantic segmentation loss (spatial cross entropy), Err2 is classification loss (cross entropy) -backWeight: loss ratio for the background -maxIter: maximum number of iterations during training -retrain: path of the pretrained model to load -nThreads: number of thread to load images -nEpochs: maximum number of epochs during training -shareGradInput: Share gradInput tensors to reduce memory usage

[OrkunYilmaz]

Thankyou for your reply! I have one last question. I am getting only FPs for the crossroads. Do you know why?

file: ./output/out7_cross.txt tp: 0 fp: 956 fn: 0 precision: 0 recall: -1 Fmeasure: 0