As show in readme and wsi_colors. there are 7 classes. However in test dataset, there are totally 6 classes.
Two class are present only in inference result, and one class only present in the test set.
It seems the label is not consistent in test dataset, in some patches white represent tissue（e.g 0e99786b-efbd-e936-5e70-bad7ef5912da_122209 [d=5.69701,x=32086,y=0,w=2916,h=2917].png） and in others it is background(e.g 0e99786b-efbd-e936-5e70-bad7ef5912da_122209 [d=5.69701,x=40836,y=0,w=2917,h=2917].png)
in this case, I wonder how I can reproduce the result of
As show in readme and wsi_colors. there are 7 classes. However in test dataset, there are totally 6 classes.
Two class are present only in inference result, and one class only present in the test set.
It seems the label is not consistent in test dataset, in some patches white represent tissue（e.g 0e99786b-efbd-e936-5e70-bad7ef5912da_122209 [d=5.69701,x=32086,y=0,w=2916,h=2917].png） and in others it is background(e.g 0e99786b-efbd-e936-5e70-bad7ef5912da_122209 [d=5.69701,x=40836,y=0,w=2917,h=2917].png)
in this case, I wonder how I can reproduce the result of
