[GitHub] some-guy1 opened a new issue #10890: [R] Is it possible to train Mxnet symbols and a GLM pretrained model as loss?

[GitBox <gi...@apache.org>]

some-guy1 opened a new issue #10890: [R] Is it possible to train Mxnet symbols and a GLM pretrained model as loss?
URL: https://github.com/apache/incubator-mxnet/issues/10890
 
 
   The idea is very similar to a GAN, but instead of two neural networks I want to use a neural network generator and a prebuilt GLM as the discriminator.
   
   Is this possible using the Mxnet symbols?
   
   Example pseudo-code:
   ```
   data <- mx.symbol.Variable('data')
   label <- mx.symbol.Variable('label')
   dconv <- mx.symbol.Deconvolution(data= data , kernel = c(3,3), stride=c(3,3), num_filter = 5)
   norm = mx.symbol.BatchNorm(data= dconv, fix_gamma=FALSE)
   act1 <- mx.symbol.LeakyReLU(data= norm, act_type = "leaky", name="act_3")
   
   #GLM that was already built and will not change
   class = abs(predict.glm(act1) - label)
   NN_Model <- mx.symbol.MakeLoss(class )
   ```
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services