Virgile Boulanger comments

Repositories
Issues
Comments

Results 14 comments of


                                            Virgile Boulanger

Shouldn't be StylePredictor and Transfer Network trained jointly ?

Hi @philipjackson and thanks for your detailed answer ! I am indeed interested in re training the StylePredictor (as well as the transfer network). But i am still a bit...

Training on FFHQ - 1024px

Got it, thanks !

Result Mismatch with the original results in paper in human faces

Hi @mu-cai, Did the above corrections lead to better structure/style swapping results on your side ?

Trained weights

Hi, thanks for sharing your work. This FFHQ trained model was trained at 256 px resolution ? For how many iterations ? Seems likes the embeddings from it does not...