Hello, first of all thanks for your excelent work.
I tried to reproduce your lsun bedroom inversion and results are not that bad. Although they are not the same as the one in the paper, it can be said satisfying.
I have pretrained StyleGAN2 model for X-ray images and it generates good X-rays. When I use this model with multi_code_inversion, the results are terrible. I run the code with following parameters. (I tried different combinations of optimization, composing_layer or z_number parameters, however results were never improved.)
python multi_code_inversion.py
--video False
--target_images images
--inversion_type StyleGAN-w+
--gan_model stylegan2_model
--optimization Adam
--image_size 256
--outputs outputs
--composing_layer 6
--z_number 30
Thanks in advance..
Hello, first of all thanks for your excelent work.
I tried to reproduce your lsun bedroom inversion and results are not that bad. Although they are not the same as the one in the paper, it can be said satisfying.
I have pretrained StyleGAN2 model for X-ray images and it generates good X-rays. When I use this model with multi_code_inversion, the results are terrible. I run the code with following parameters. (I tried different combinations of optimization, composing_layer or z_number parameters, however results were never improved.)
Thanks in advance..