not good inversion as seen in papers

[adeelnazar]

why above picture of donald trump inversion is bad and another is good ??

[wty-ustc]

I guess he wasn't correctly aligned.

[adeelnazar]

I guess he wasn't correctly aligned.

hi wty-ustc

do you have a inversion.pt model file trained with celeba-HQ dataset if you have kindly give me a link

thanks

[adeelnazar]

I guess he wasn't correctly aligned.

Please send me inversion.pt model file trained with celeba HQ dataset please give me a link

[wty-ustc]

Sorry, there is no pre-trained model on this dataset, you can train one yourself. However, my personal experience is that the model pre-trained on FFHQ performs well enough.

[adeelnazar]

Can you guide me how to bestv inversion for hd face like Hollywood actor face Sent from Yahoo Mail on Android On Fri, Jan 20, 2023 at 1:22 PM, ***@***.***> wrote: Sorry, there is no pre-trained model on this dataset, you can train one yourself. However, my personal experience is that the model pre-trained on FFHQ performs well enough. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

[wty-ustc]

Please refer to #22. You should first use the script to align the image to be inverted and then use the aligned image for the inversion.

[adeelnazar]

Sorry, there is no pre-trained model on this dataset, you can train one yourself. However, my personal experience is that the model pre-trained on FFHQ performs well enough.

Can we train combine data of both ffhq and celeba HQ and train is it possible to give better results ...please guide me