Sacha
Sacha
No, I haven't sorry.
These were merged in. Thanks.
Nevermind, it's still relevant for CurricularFace.
If training precision gets to 99%, you have overfitted the data. It'll probably eventually hit 80-90% depending on depth if you leave it training much longer with lower learning rates,...
Yeah that log looks wrong. You definitely should never get 100% training accuracy. Even close to it is bad. Your model will probably be bogus if you get 100% trained...
This happens in the last bottleneck only, where depth and channels are both the same and stride is not set (so defaults to 1), right? As far as I know...
I think for the data augmentation, you need to train it longer. You can also resume from baseline and train from there instead (which is what I do). I always...
@cavalleria my results wouldn't be compatible as I use a different face detector and dataset(s). I start with the fully trained baseline backbone and retrain with LR=0.01 + augmentations turned...
LFW should have a perfect score (99.83+%) very early on.. like after first LR drop. Edit: You mean on masks? You should be getting around 99.5% after first LR drop...
It depends on the distribution of the test data as to how well the augmentation will score. Ultimately, if you have perfect frontal and high quality photos for train and...