search

yu4u / age-gender-estimation

Keras implementation of a CNN network for age and gender estimation
MIT License
1.47k stars 503 forks source link

MAE(7.436) and e-error(0.543) compared with References[2]'s best score MAE(3.252) and e-error(0.282) #22

Open zzjing83 opened 6 years ago

zzjing83 commented 6 years ago

Hi Yusuke. Thanks your repo firstly. It helps me a lot. My work following your repo is as below: 1, I cleaned the imdb dataset manually and converted them to imdb_zz.mat. 2, Training on imdb_zz.mat. --aug Ture, depth 16 width 8 as default. num_epochs 30. 3, the minimum val_loss is 3.49 and the weights stored as weights.25-3.49.hdf5. 4, Using weights.25-3.49.hdf5 to predict images. I test on LAP dataset as References[2] mentioned. However, my score is much worse than References[2] as discribled in title and it predicts young people(age<15) badly.

i am looking forward to your reply and any advice is welcomed. Thank you.

yu4u commented 6 years ago

Thank you for testing this project on the LAP dataset! Please note that this project is not intended to reproduce the results of the papers [1, 2] (of cause I'm interested in doing that). Anyway, I think the most important thing to improve the accuracy is to fine-tune the model using the LAP training dataset. Did you try that? Indeed it is not trivial because the LAP dataset is for a single task of age estimation while age and gender are simultaneously estimated on the model of this project.

Riolite5 commented 5 years ago

Hello everyone, what is the difference between loss and val_loss ?