about age and gender branch, age precision is very low

[hewumars]

enviroments:

• ubuntu 16.04
• pip install -r requirement2.1.txt

commad

1. age and gender python test.py -g -white -se -i data/Age_Gender/ -m model/c3ae_model_v2_fp16_white_se_132_4.208622-0.973
2. only age python test.py -se -i data/Age_Gender -m model/c3ae_imdb_v89.h5

result compare

1. age and gender result
2. only age result

I changed the single image test to batch images test for average precision.but the same test data but the results are very different.

Is it the parameters configuration problem or the model problem?

[StevenBanama]

I have had test it in our inner dataset, which worked normally.But now I have changed my job and no GPU at hand.Maybe you can retrain or finetune it.