No GPU devices found

[liuliuliu11]

hi, when i carry out convert_weight.py, it occured a error:

[zengxianyu]

It's probably because your tf dosen't have gpu support

[zengxianyu]

I uploaded the models I converted

https://maildluteducn-my.sharepoint.com/:u:/g/personal/zengyu_mail_dlut_edu_cn/EUoHD5628IxOjPb8jbWKSWsBoqOLUTsYeLXjXe9i5ahxvQ?e=Qs7SuX

[cyq373]

I uploaded the models I converted

https://maildluteducn-my.sharepoint.com/:u:/g/personal/zengyu_mail_dlut_edu_cn/EUoHD5628IxOjPb8jbWKSWsBoqOLUTsYeLXjXe9i5ahxvQ?e=Qs7SuX

Thank you!

[Zeeshan75]

Hi, Thanks for sharing the models, I have a little doubt, the model which you have shared are half of the .pkl files. i,e .pt files are half of the .pkl files, the files which you have shared are only encoder? or do we have even a decoder file?

what is the config-a and config-f refer to and why you have converted only config-f files?

[lww28]

@zengxianyu hii, thanks for sharing the converted weights of the model. However, there are a few missing keys. Have you ever tried generate.py, how do you solve this problem? I tried to use strict=False g_ema.load_state_dict(checkpoint["g_ema"],strict=False) to generate the image but the quality doesn't look good. Do you have any thoughts on it? Thank you!

RuntimeError: Error(s) in loading state_dict for Generator: Missing key(s) in state_dict: "convs.0.conv.blur.kernel", "convs.2.conv.blur.kernel", "convs.4.conv.blur.kernel", "convs.6.conv.blur.kernel", "convs.8.conv.blur.kernel", "convs.10.conv.blur.kernel", "convs.12.conv.blur.kernel", "convs.14.conv.blur.kernel", "to_rgbs.0.upsample.kernel", "to_rgbs.1.upsample.kernel", "to_rgbs.2.upsample.kernel", "to_rgbs.3.upsample.kernel", "to_rgbs.4.upsample.kernel", "to_rgbs.5.upsample.kernel", "to_rgbs.6.upsample.kernel", "to_rgbs.7.upsample.kernel".

[TingHLiao]

@zengxianyu hii, thanks for sharing the converted weights of the model. However, there are a few missing keys. Have you ever tried generate.py, how do you solve this problem? I tried to use strict=False g_ema.load_state_dict(checkpoint["g_ema"],strict=False) to generate the image but the quality doesn't look good. Do you have any thoughts on it? Thank you!

RuntimeError: Error(s) in loading state_dict for Generator: Missing key(s) in state_dict: "convs.0.conv.blur.kernel", "convs.2.conv.blur.kernel", "convs.4.conv.blur.kernel", "convs.6.conv.blur.kernel", "convs.8.conv.blur.kernel", "convs.10.conv.blur.kernel", "convs.12.conv.blur.kernel", "convs.14.conv.blur.kernel", "to_rgbs.0.upsample.kernel", "to_rgbs.1.upsample.kernel", "to_rgbs.2.upsample.kernel", "to_rgbs.3.upsample.kernel", "to_rgbs.4.upsample.kernel", "to_rgbs.5.upsample.kernel", "to_rgbs.6.upsample.kernel", "to_rgbs.7.upsample.kernel".

Hi @wenWern, I met the same issue, do you solve this? Thanks.