search

yuval-alaluf / stylegan3-editing

Official Implementation of "Third Time's the Charm? Image and Video Editing with StyleGAN3" (AIM ECCVW 2022) https://arxiv.org/abs/2201.13433
https://yuval-alaluf.github.io/stylegan3-editing/
MIT License
654 stars 73 forks source link

Some thing weird with FFHQ-00000.png #33

Closed zhongtao93 closed 2 years ago

zhongtao93 commented 2 years ago

Hi, thank you for sharing source code! I test 00000.png in FFHQ dataset, but it seems not very well. image I get this results by cmd below： python inversion/scripts/inference_editing.py \ --output_path /DATA/jupyter/tmp_data/FFHQ-1024x1024-beyoung/ \ --checkpoint_path pretrained_models/restyle_e4e_ffhq.pt \ --data_path /DATA/jupyter/tmp_data/FFHQ-1024x1024/ \ --test_batch_size 4 \ --test_workers 4 \ --n_iters_per_batch 3 \ --edit_directions "[age]" \ --factor_ranges "[(-5_5)]"

Hope for your reply~

yuval-alaluf commented 2 years ago

Does your --output_path point to the directory where your inference results were saved to?

zhongtao93 commented 2 years ago

Yes. But it seems to be a special case. I tested dozens of other FFHQ pictures, but this didn't happen