search

researchmm / AOT-GAN-for-Inpainting

[TVCG'2023] AOT-GAN for High-Resolution Image Inpainting (codebase for image inpainting)
https://arxiv.org/abs/2104.01431
Apache License 2.0
403 stars 64 forks source link

Unable to train #19

Closed s8088899 closed 1 month ago

s8088899 commented 1 month ago

I can't use my own training set, I prepared some 512*512 images for training, but my progress bar keeps staying at 0, like this:

[**] create folder ../experiments/aotgan_places2_pconv512 UserWarning: The parameter 'pretrained' is deprecated since 0.13 and will be removed in 0.15, please use 'weights' instead. warnings.warn( UserWarning: Arguments other than a weight enum or None for 'weights' are deprecated since 0.13 and will be removed in 0.15. The current behavior is equivalent to passing weights=VGG19_Weights.IMAGENET1K_V1. You can also use weights=VGG19_Weights.DEFAULT to get the most up-to-date weights. warnings.warn(msg) 0%| | 0/1000000 [00:00<?, ?it/s]

s8088899 commented 1 month ago

I have solved the problem, but the model gives an error when saving: AttributeError: 'InpaintGenerator' object has no attribute 'module'