zyqz97 / gp-nerf Goto Github PK

Thanks for your great work.

I have a confusion about the setting of foreground and background. The default bg_nerf is off in the code, which seems inconsistent with the paper.
Therefore I set bg_nerf=true, but I found that it did not improve the results. Is there a critical step I have overlooked?

Looking forward to your reply.
Thanks again.

Hi, thanks for your great work! When I try to run your code, I encounter the following problem. Could you mind tell me how to fix it?

Traceback (most recent call last):
File "gp_nerf/train.py", line 35, in
main(_get_train_opts())
File "/home/zyc/anaconda3/envs/k-planes/lib/python3.8/site-packages/torch/distributed/elastic/multiprocessing/errors/init.py", line 346, in wrapper
return f(*args, **kwargs)
File "gp_nerf/train.py", line 21, in main
from gp_nerf.runner_gpnerf import Runner
File "/home/zyc/Desktop/GP-NeRF-main/./gp_nerf/runner_gpnerf.py", line 28, in
from mega_nerf.datasets.filesystem_dataset import FilesystemDataset
File "/home/zyc/Desktop/GP-NeRF-main/./mega_nerf/datasets/filesystem_dataset.py", line 24, in
class FilesystemDataset(Dataset):
File "/home/zyc/Desktop/GP-NeRF-main/./mega_nerf/datasets/filesystem_dataset.py", line 310, in FilesystemDataset
img_indices_dtype,desired_chunks=1000) -> List[Future[None]]:
TypeError: 'type' object is not subscriptable

Hello, pls what is the version of numpy, I got an error when I reproduced the code. Specifically, I used the latest numpy-1.26.2 as well as mega-nerf's numpy-1.21.2 and both got an error. Hope to hear back from you.

Thank you very much for sharing your excellent work.
May I inquire if you could provide the pretrained models? In any case, thanks for your help.

When I was 95% trained, I reported this mistake, how to solve it?

Hi, thank you for releasing your code to accompany your paper. Pardon me if I've missed something, but I don't seem to see a LICENSE associated with this repository. Would you please be able to add one? Both Mega-NeRF and torch-ngp use MIT license which might also make sense to use here. Thank you!

Hello, thanks for the great job.

I wanted to ask if this approach uses submodules similarly to the Mega-NeRF, if yes what replaced the merge_submodules.py since I noticed that both approaches uses similar training phase.

Thanks in advance.

你好作者，感谢你开源了论文中的代码，我想问一下如果我想结果可视化的话我应该怎么做？给定几个位姿渲染几张图片就行

Thanks for your work.

When using Mega-NeRF and GP-NeRF, I have noticed that the validation set might have been inadvertently included in the training set (in this issue). It seems that this issue has also been carried over to GP-NeRF.

I'm not certain if I may have overlooked an important part of the code, but if there are any errors, please feel free to point out my mistakes.