A problem about evalution

[CodeH0922]

Excuse me, when I attempt to train or test the model, there is always an error:"AttributeError: 'ConcatDataset' object has no attribute 'evaluate'". So how can I solve this problem?

[Zeal24]

Have you solved this problem?

[CodeH0922]

Have you solved this problem?

Yes.

[Zeal24]

Have you solved this problem?

Yes.

how do you solve this problem? I need your help,tanks a lot!

[CodeH0922]

The file to be run is not “test.py”, but “bonai_test.py”. It can be executed if the path is set correctly

Have you solved this problem?

Yes.

how do you solve this problem? I need your help,tanks a lot!

[Zeal24]

The file to be run is not “test.py”, but “bonai_test.py”. It can be executed if the path is set correctly

Have you solved this problem?

Yes.

how do you solve this problem? I need your help,tanks a lot!

The file to be run is not “test.py”, but “bonai_test.py”. It can be executed if the path is set correctly

Have you solved this problem?

Yes.

how do you solve this problem? I need your help,tanks a lot!

sorry to bother you. I meet the problem when I train this model, I dont konw where to set the path

[smallpotato000]

In datasets/dataset_wrappers.py, add evaluate() in class ConcatDataset(_ConcatDataset).
I did this by backporting from mmDetection 2.4.0..
Please note that BONAI is a fork of mmDetection, between 2.3.0 and 2.4.0. Code in this repo is imcomplete, and does not work out of the box. I have got it working by a few quick & dirty fix. Will share my work in about 2~3 weeks, after some clean up.