search

GilLevi / AgeGenderDeepLearning

628 stars 284 forks source link

Tensorflow models #19

Closed red010182 closed 6 years ago

red010182 commented 7 years ago

This is not a bug issue.

In https://github.com/dpressel/rude-carnie, it seems use a tensorflow model provided from this repo, in his example it's located at AgeGenderDeepLearning/Folds/tf/age_test_fold_is_1/run-20854, but I can't find it anywhere in both repositories.

Are there tensorflow version models I just missed? Or, should I train it in tensorflow myself to get the models?

Really appreciate for sharing this work.

GilLevi commented 7 years ago

Hi again @red010182,

I'm afraid I don't know. I think you should ask dpressel, the author of the tensorflow repo, he will probably know the answer.

Best, Gil

RBirkeland commented 6 years ago

I belive you found the answer in another thread. dpressel/rude-carnie#37

dpressel commented 6 years ago

Sorry for any confusion. It uses the setup from this repository to create the folds and the rude-carnie checkpoints happened to exist in the directory for the example.

The checkpoints in question are trained by using rude-carnie, which is a tensorflow implementation. The answer in the other ticket has to do with the checkpoint type. rude-carnie supports two different approaches for age/gender detection. The default method of training/evaluation is to use Gil's model architecture (re-implemented in tensorflow), but it also supports fine-tuning a pre-trained imagenet inception model. It happens that the checkpoint that I uploaded was for fine-tuned inception, and so the user had to pass in the --model_type inception for it to run.