-
Hi, I was experimenting with AsoftMax and mobilenet for Cifar10, but it seems the training results are very erratic and can't be explained when we use the Asoftmax layer. Following are the scripts and…
-
# SphereFace: Deep Hypersphere Embedding for Face Recognition #
- Author: Weiyang Liu, Yandong Wen, Zhiding Yu, Ming Li, Bhiksha Raj, Le Song
- Origin: https://arxiv.org/abs/1704.08063
- Related:…
-
也就是在model中的 last_cont 是表示什么含义呢？对应于论文上的哪一部分呢？ 根据上面的判断，只会在train 的最后一步，数据不够的一个batch的时候触发，为什么要单独处理呢？
-
**how to predict the class probability? when I set the output to arcface output (softmax) w.r.t number of class, I got an error when run this model = Model(inputs=model.input[0], outputs=model.layers[…
-
Since m=3 hasn't been implemented, will you please provide the formula for that.
ghost updated
5 years ago
-
what is the meaning of 'f' and 'ff' ? 0.5???
-
Hi,
I have tensorflow dataset
(ds_train,ds_test), ds_info = tfds.load('grains', split=["train","val"], with_info=True, as_supervised=True)
where ds_train consist of batch of image and label.
ho…
-
i want to use the trained model to predict the class of a test image without sending the label, is it possible?
currently, the implementation required sending the Keras "predict" function a tuble …
-
Hi guys,
I tested A-Softmax loss on CASIA dataset and it really does well, reaching a ~99% performance on lfw without careful tuning. But when I switch to use the MS-Celeb-1M datasets for training(…
-
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…