-
Hi wy1iu, thanks for your sharing codes. I try to train the webface_casia datasets, the lossvalue convergence. but I train the ms1m datasets ,the lossvalue divergence. The params setting as follow:
i…
-
In the forward function, why mulitiply the norm of feature after output ?
`output = (one_hot * (phi_theta - cos_theta) / (1 + self.lamb)) + cos_theta
output *= NormOfFeature.view(-1, 1)
`
-
hello, in your code, I have seen several line codes like this:
if mbatch
-
Dear author, I used the default configuration hope to reproduce the resutls, the acc alwyas show 0 after 2000 Batch. Is there anything wrong?
I only have 2 GPUs, so I changed the visible device to 0,…
-
-
Hello,
In your ReadMe.md, in the section "Pretrained Models", you say that "You can use $INSIGHTFACE/src/eval/verification.py to test all the pre-trained models."
When I look inside this verificat…
-
(1) when train loss type 2 with sphereface net, there will be a shape error, as follows; but use other nets, it's ok. The error occurs in the sphereface_net.init_weights when infer_shape, but I can't…
-
根據這兩篇文章([link](https://zhuanlan.zhihu.com/p/34717205))([link](https://blog.csdn.net/fuwenyan/article/details/79657738)),分別實作 :
- [ ] L-softmax ([link](https://arxiv.org/pdf/1612.02295.pdf))
- [x] A…
-
I try to transfer the sphere face model (https://github.com/clcarwin/sphereface_pytorch) from pytorch to caffe2. But the code remind me such error.
RuntimeError: ONNX export failed: Couldn't exp…
-
How do you use it?