Prime Cai

I'm wondering if anyone attepted to fix this issue and how's the performance? Many thanks.

Hi! I have not tried it yet, but I would assume the mesh quality will be relatively similar to pi-GAN. I will perform some experiments on this and let you...

Hi, I tested it with a 12GB GPU. Perhaps try decreasing `max_batch_size`? The rendered images should be saved to `output_dir`, so there is no need to run any additional script....

Hi, The output resolution can be changed by specifying a higher "img_size" in "render_options": https://github.com/primecai/Pix2NeRF/blob/ee036fe5592c65dd6cf57ee8486a242ce4e25c08/hybrid_optimization.py#L63 Note that we train the model with 64x64 images, so naively setting a resolution very...

Hi, Many thanks for the quick reply. So for the case of Places2 pre-trained model, the output 'iters' in 'train.log' should actually be 4 * 10 * 660k = 26400k...

Many thanks again. May I also ask from what point the model starts to inpaint some very reasonable outputs? Best regards, Shengqu

Hi, The "cleanup" function is specific to a cluster I am using. Please feel free to comment it out! It has no effect on training. Best regards, Shengqu

Are you sure the other parts of the code are unchanged, and that you are loading the datasets correctly? The time it takes to reach 6 epochs is definitely not...

Hi, I do not think training on 1280x720 images with noisy backgrounds will work without any tuning. There are two reasons: 1. Our backbone pi-GAN is very memory consuming to...

Hi, Yes, it will be very interesting to train a feed-forward framework for EG3D by combining the ideas. And I may not fully understand your second question, do you mean...