Errors related to the generated image about stylegan2-ada-pytorch

[ramnoob]

@wry123456
I would be very happy if you could share how to do this.

[whyydsforever]

@wry123456 I would be very happy if you could share how to do this.
if your dataset is gratscale, you can delete the last dimesion of image by
adding img = np.squeeze(img, axis=3) in line 121
it will solve this problem

[FwiffoSnork]

Is your pickle trained on greyscale or RGB images?

[wry123456]

您的泡菜是否经过了灰度或RGB图像训练？

thanks for your hints. My training set data is grayscale images, and I modified the way the images are saved in the generation code. It can be run.

[Elkortya]

FWIW, I also has to change the "RGB" string in the (original) line 121 to "L" for it to work.