Promblems of loading pre-trained models for testing

[mingju-c]

When I try to use pre-trained models for testing, i.e. run bash scripts/segpn_eval.sh. There will always be errors as shown in the figure below , but when I train your method under the few-shot setting, i.e. run bash scripts/segpn.sh, there will not appear similar errors, want to know why?
Hope for some explanations, thanks a lot!!!

[Wzy-lab]

I met the same problem as you. Have you solved it?