Open abdikaiym01 opened 2 years ago
I only tried the vanilla mixup augment method very earlier, and the results not good. I think it's not compatible with margin loss functions like ArcFace
:
0.5
mixup, [0, 0, ..., 0.5, ..., 0.5, 0, ...]
ArcFace
makes a margin for truth predictions, like 0.8 -> 0.4
, 0.5 -> 0.1
. This makes them pretty low.Some basic results using EfficientNetV2B0 + MS1MV3:
lfw: 0.9975, cfp_fp: 0.975714, agedb_30: 0.976333
lfw: 0.997833, cfp_fp: 0.973571, agedb_30: 0.974333
Hi, have you tested the technique mixup (Mixup: Beyond Empirical Risk Minimization) for the different datasets? Is with this technique models can show better performance on dataset IJBC?