Difference between pre-trained models with suffix '-devel' and '-test'

[Mu-Y]

Hi,

I am running classify.py with different pre-trained models. I read from the wiki that when suffix is omitted, the '-test' will be used as default. So I'm using the following two commands:

python classify.py -i GE09-devel -m GE09 -o GE09-devel
python classify.py -i GE09-devel -m GE09-devel -o GE09-devel-usingDevelModel

The f1 scores are quite different though(the second one is much lower than the first one). Is this expected? Can you kindly explain the difference between the two pre-trained models: 'GE09' and 'GE09-devel'? If possible, can you tell me which model is used for your reported f1 scores here? Appreciate your help!

Thanks!