NN shape/design details

Home Forums DeepFaceLab Development NN shape/design details

Viewing 1 post (of 1 total)
  • Author
    Posts
  • #8842
    whiterabbitobj
    Participant

      I am trying to find out more details about the shape of SAEHD models and the relation of dimensions selected when beginning a model.

      Specifically, what is the relationship of AE Dims, Encoder, Decoder, and output resolution. What does each of these actually do to the network design?

      I tried reading the code but the model.py code is pretty dense and I’m having trouble finding the network initialization.

      Thanks.

    Viewing 1 post (of 1 total)
    • You must be logged in to reply to this topic.