modadao

train time

1

Hello, how long does it take you to train a model (such as brandenburg_gate), and what hardware configuration do you use for training?

test

1

1. As for testing, for example, 10 images were used to test the Brandenburg Gate model. Why are these 10 tested images almost free of occlusion? Can I use more...

tsv file

1

There are only 6 tsv files. If I want to train files other than these 6 tsv files, what should I do? Do you want to manually create a tsv...

Why does the calculation of loss need to be multiplied by a coefficient of 0.5? Is it because fine and coat account for half respectively? ret['c_l'] = 0.5 * ((inputs['rgb_coarse']-targets)**2).mean()...

colmap data

1

If it is a colmap dataset, cuda memory will out of memory?

How should I test metrics（such as dtu dataset or real dataset）? In Wandb, I only see the PSNR of val.I hope you can tell me，thank you.

Images instead of videos

4

Can I use the already captured images and obtain JSON files without using videos

I want to test, but the link to Google Drive is not available： Please download the weights from [GoogleDrive], and put it into the [weights/](https://github.com/Mukosame/AODA/blob/main/weights) folder.

core dumpe

1

 How to solve the above problems when I train stage 1?

Extract the mask of the entire image

4

Hello Xingyu, may I ask how to extract the mask of the entire image based on the pre trained model you provided?