Are there any code examples for the application of this kind of VAE to semi-supervised classification, as described in section 4.3 of the Gumbel-Softmax paper?
I understand the technical details in the appendix, but I'm having trouble defining the right loss function for the model depicted in Figure 6. Any help would be much appreciated!
Are there any code examples for the application of this kind of VAE to semi-supervised classification, as described in section 4.3 of the Gumbel-Softmax paper?
I understand the technical details in the appendix, but I'm having trouble defining the right loss function for the model depicted in Figure 6. Any help would be much appreciated!