Why does the test and evaluation shows different results?

[jhkim0759]

I'm training ROMP model with your code It saves every validation process but, I test the results with "python -m romp.test" code I got the different result as evaluation I think they are same process, do you know why ? Or did I do something wrong?

[Arthur151]

Yes, i observe the same thing, which is quite weird. Haven't figured out the reason for now.

[jhkim0759]

Thank you. I'll tell you when I find out why.

[jhkim0759]

I found it!! its because of the 'match_preds_to_gts_for_supervision' argument. In evaluation config the 'match_preds_to_gts_for_supervision' was set to false Thank u.