Unable to reproduce the performance on ConSep Dataset

[PrashastiSachan]

Hi,

Thanks for sharing your work. I am using pytorch branch to train the model. However, it is not performing well on validation data during training. The settings are same as your code in hover/opt.py and config.py Am I missing any step to run before inference? Here is the training metrics when training with the pretrained "npz" file

Let me know how can I correct it to reproduce your results

[VolodymyrChapman]

I may have answered this in another issue: issue 220 Please confirm that you have used exactly the same learning process as stated in the original paper. By default, this HoVerNet GitHub has somewhat different parameters.

[zhouyi-git]

Why can't I find the value of the three output indicators of the test, for example, Dice=0.85, where is the output?

[hjtalent2023]

Hello! I have met a similarly question with you. like flow: ------valid-np_acc : 0.92795 ------valid-np_dice : 0.73184 ------valid-tp_dice_0 : 0.94518 ------valid-tp_dice_1 : 0.02475 ------valid-tp_dice_2 : 0.49475 ------valid-tp_dice_3 : 0.00000 ------valid-tp_dice_4 : 0.61169 ------valid-hv_mse : 0.06671 the dice_3 is always 0. I use the Consep dataset,and set the nr_types are 5. Does anybody can help me?