-
Original Post by @neuraldomi: https://groups.google.com/forum/#!topic/cmu-openface/lhT5Cwjx8XQ
> My current setup is that I train a DNN on a CUDA server (also classifiers I create with CUDA) and I wo…
bamos updated
8 years ago
-
```
training(master*)$ ./main.lua -data ~/openface/data/casia-facescrub/dlib-affine-224-split/ -imgDim 224 -batchSize 20 -nDonkeys -1
==> doing epoch on training data:
==> online epoch # 1
/home/bam…
bamos updated
8 years ago
-
Hi!
I want to train a model in DIGITS with a contrastive loss function, but I get an error telling me to add a loss layer. Is it possible to use any other loss function different from "SoftmaxWithLos…
-
I use the CASIA_train_test.prototxt with my dataset(2000person and 50 image for each total 10W+).
I want to training deepid network using CASIA_train_test.prototxt. Only change the "ip1" layer output …
vzvzx updated
9 years ago
-
hi yang,
many thanks for your codes, it is clean and works well.
but i tested on the new released casia-webface dataset, facex fails to generate meaningful landmarks on a large portion of images, ma…
-
Hi, why the features stored in the matlab matrix is 256 dimensions? but fc1-layer in the network's output number is 512. I cannot understand which layer's output treat as the terminal feature!
-
I designed a network using caffe to do face verification. I chose CONTRASTIVE LOSS to optimize the net. But I met a problem :
The loss can not descend to a reasonable level. I have tried many strate…
-
Color per vertex meshes are not supported and should be. Offending meshes can be found in the CASIA 3D dataset (they have to be converted from WRL to OBJ using MeshLab).