Clorr
Clorr
I found an answer here: https://github.com/keras-team/keras/issues/3303
I'm playing a bit with that `lr` value. I tried: - 5e-3 gives erratic results, and leads to complete mess up - 5e-4 seems fine, still a bit erratic, but...
Ah nice if we can use `fit`. I thought it was not possible because of the encoder that is shared. Having `fit` is also nice as we can use some...
@Ganonmaster did you see that ? What do you think about it ?
Great then ! I'm thinking about a plugin architecture, already for alignment. It won't be very sophisticated but at least it will enable us to have different behaviors.
Someone posted this on Reddit: https://www.reddit.com/r/deepfakes/comments/7n4w9p/comparison_of_loss_functions_model_decoder/ It adds some layer to the model. What's your opinion on this guys?
@shaoanlu if you think that providing a "generic" src data set could help make generic models, it would already be a major impprovment... @Ganonmaster my first tries with the code...
Hi @bryanlyon and thanks for your feedback. I'm not sure to understand all of your answer as english is not my first language and maybe i missed things... My idea...
@JarbasAl Shaoanlu has also tried [face landmarks as source here (see 6.)](https://github.com/shaoanlu/faceswap-GAN/tree/master/notes)
Hello @curry684 Today, we encountered the same error. We had a dataTable using MongoDBAdapter and a `filter` entry (see code below). The search was not working, but with the code...