Hello, I'm sorry to bother you.
I have encountered some problems in using this network. When segment, I want to use the categoryID of cityscapes for training. I put the cityscapes in the colors folder TRID in py TO LABEL = labels cityscape seg. Gettrainid2label() is modified to TRID TO LABEL = labels cityscape seg.getcategory2label()。 However, the trained model is still 19 classes, not 7.
I want to know how to correctly modify the training semantic labels.
I hope to get your answer. Thank you very muchs
Hello, I'm sorry to bother you. I have encountered some problems in using this network. When segment, I want to use the categoryID of cityscapes for training. I put the cityscapes in the colors folder TRID in py TO LABEL = labels cityscape seg. Gettrainid2label() is modified to TRID TO LABEL = labels cityscape seg.getcategory2label()。 However, the trained model is still 19 classes, not 7. I want to know how to correctly modify the training semantic labels. I hope to get your answer. Thank you very muchs