I'm having a problem reproducing MISSFormer on Synapase dataset

[LebronMX-A]

I'm having a problem reproducing MISSFormer on Synapase dataset, when testing it, I get a situation where the dice of certain classes are 0, similar to the following. How can I solve this?

0it [00:00, ?it/s]idx 0 case case0008 mean_dice 0.412015 mean_hd95 21.133825

1it [01:24, 84.54s/it]idx 1 case case0022 mean_dice 0.678225 mean_hd95 26.268258
2it [02:18, 66.26s/it]idx 2 case case0038 mean_dice 0.627728 mean_hd95 6.501879
3it [03:19, 64.16s/it]idx 3 case case0036 mean_dice 0.663994 mean_hd95 16.518772
4it [05:22, 87.21s/it]idx 4 case case0032 mean_dice 0.647384 mean_hd95 16.401192
5it [06:58, 90.41s/it]idx 5 case case0002 mean_dice 0.664552 mean_hd95 3.900517
6it [08:30, 90.88s/it]idx 6 case case0029 mean_dice 0.662251 mean_hd95 5.307463
7it [09:28, 80.32s/it]idx 7 case case0003 mean_dice 0.539956 mean_hd95 41.322588
8it [11:34, 94.68s/it]idx 8 case case0001 mean_dice 0.652544 mean_hd95 22.512179
9it [13:08, 94.53s/it]idx 9 case case0004 mean_dice 0.631242 mean_hd95 16.877601
10it [14:36, 92.52s/it]idx 10 case case0025 mean_dice 0.524473 mean_hd95 26.227055
11it [15:29, 80.41s/it]idx 11 case case0035 mean_dice 0.657376 mean_hd95 2.458756
12it [16:26, 82.21s/it]
Mean class 1 mean_dice 0.824931 mean_hd95 19.636354
Mean class 2 mean_dice 0.000000 mean_hd95 0.000000
Mean class 3 mean_dice 0.825062 mean_hd95 8.764452
Mean class 4 mean_dice 0.760876 mean_hd95 18.267999
Mean class 5 mean_dice 0.921958 mean_hd95 22.872335
Mean class 6 mean_dice 0.000000 mean_hd95 0.000000
Mean class 7 mean_dice 0.865811 mean_hd95 38.525610
Mean class 8 mean_dice 0.709188 mean_hd95 28.886639
Testing performance in best val model: mean_dice : 0.613478 mean_hd95 : 17.119174

[10dutel]

hello, i have the same question, could you tell me how to solve this problem?