-
I am experiencing a significant discrepancy in the logits and probabilities output when running the same TensorFlow.js model on Android and iOS. The model, a self-trained MobileNetV3 (large) from PyTo…
-
File "D:/project/pydemo/Angular-Penalty-Softmax-Losses-Pytorch/train_fMNIST.py", line 47, in main
plot(am_embeds, am_labels, fig_path='./figs/{}.png'.format(loss_type))
File "D:\project\pyde…
-
Similar to #27 but for decoding. We should have built-in defaults for common decoding schemes, like softmax, first spike, population, angular, etc.
Any others @cpehle ?
-
## 🚀 Feature
Adding support for Additive Angular margin loss to PyTorch as described by *Deng et.al* in *ArcFace: Additive Angular Margin Loss for Deep Face Recognition*
## Motivation
Additive …
-
Hi @AlexeyAB ,
I read in multiple places that Face Recognition community replacing traditional Softmax with Margin Softmax, Angular Margin, Cosine Margin and and getting astounding improvement in a…
-
你好,我想跟您的paper: Angular Softmax for Short-Duration Text-independent Speaker Verification的实验对其,但是我没有NIST SRE corpus这个数据库,我在网上找了也不全,您能分享一下这个数据库吗?
-
hello,
there are some paper normalize the features before input Softmax layer, and I think it's make sense because
(1)it will make the feature space more dense which suit for hyperplane to do classi…
-
when using cosface / arcface as metric, is it right to train the model with margin and remove the margin when using the model to predict?
For example, if I have a trained model with [arcface]. Shou…
-
hello, I am a bit of confuse about reading this code about calculating A-softloss
-
Hi, @davidsandberg:
Recent paper [SphereFace: Deep Hypersphere Embedding for Face Recognition](https://arxiv.org/abs/1704.08063) introduces a new novel LOSS definition to the face recognition tra…