search

yuval-alaluf / SAM

Official Implementation for "Only a Matter of Style: Age Transformation Using a Style-Based Regression Model" (SIGGRAPH 2021) https://arxiv.org/abs/2102.02754
https://yuval-alaluf.github.io/SAM/
MIT License
632 stars 151 forks source link

how to finetune the vgg age (DEX) ? #59

Closed kuro1730 closed 1 year ago

kuro1730 commented 1 year ago

is the dex that use for training have a limit of race (eg.asian) from the dataset that use to train it ?. if it's have some limit with that . how can i finetune it?

i have plan to use AAF (All-Age-face) Dataset to finetune it . thanks for nice work!

yuval-alaluf commented 1 year ago

I don't have the code that we used for training the age classifier. But it was basically a simple regression training scheme that normalized the ages to be between 0 and 1 and computed the MSELoss between the predicted and true ages. I believe everything else was standard. The model architecture is provided in the repo so all you need to do is write the training script around this model.