1. Is the dataset for the private test the same as the dataset for the public test?
2. How do we get feedback on the private test?
3. What is the experimental environment of the private test (both software and hardware environment), do we need to configure it?

当我运行 python train.py -c ./cfgs/vcrnet.yaml 时

报错信息是这样的 ：ImportError cannot import name 'ball_query_ext' from 'mm3d_pn2.ops.ball_query'

下面是 'mm3d_pn2.ops.ball_query 这个文件夹中的内容，里面确实没有 'ball_query_ext' 这个文件

utils metrics mm3d_pn2 ops ball_query src ball_query.cpp ball_query_cuda.cu init.py ball_query.py

在 ball_query.py 这个文件里面有这么一条导入语句：

import torch
from torch.autograd import Function

from . import ball_query_ext  // 在当前目录中没有这个模块文件，请问这个错误该咋解决呢？


class BallQuery(Function):
    """Ball Query.

    Find nearby points in spherical space.
    """

    @staticmethod
    def forward(ctx, min_radius: float, max_radius: float, sample_num: int,
                xyz: torch.Tensor, center_xyz: torch.Tensor) -> torch.Tensor:
            ................
            ...................
            .................

from . import ball_query_ext // 在当前目录中没有这个 ball_query_ext.py 模块文件，请问这个错误该咋解决呢？

Hello, I tried to test the code for point cloud registration in Benchmark, but I found that the GPU utilization for these methods is very low (jumping between 0% and 20%, and the training process is also very slow), I used multiple GPUs, do you use a single one for the training process or? Can you provide us with your hardware environment and the approximate time required for a full training or an epoch? I'm not sure if it's my environment or code problem, I'd like to get your help.

Two test environments: a). cuda10.2, GPU 2080Ti, pytorch1.5.0; b). cuda11.1, GPU 3090, pytorch1.7.0.

After reading your paper about VRCNet, I found this net is based on a prior pth. But I cannot find best_cd_t_network.pth in this repo. I found pretrained_vrcnet_2048.pth in VRCNet repo, but it seemed not working... help me...

Hello, I have three questions about this competition.

1. Will it be illegal to use the pre-trained weight from other methods, like "PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers"? The result could be better if I train the network based on the pre-trained PCN/C3D weight.
2. Is the submission on the competition website still effective during the private submission stage?
3. Can we only give the private submission without showing it on the leaderboard?

Thank you very much!

I don't know if it is correct, I think there may be some bugs in the ranking order of the point cloud registration leaderboard? It looks like it’s currently sorted by submission time, not by some kind of metric.

Hi,

thanks for releasing the code and dataset. However, on the paper you provide results at different resolutions (2.048 - 4.096 - 8.192 16.384), which are not present in the current released dataset. Why have they been removed ?

Thanks Riccardo

Hi, @paul007pl ,

Thanks for releasing the code for the benchmark and baseline models. Do you also provide the trained models (checkpoints) for baseline methods (e.g. [1] PCN; [2] ECG; [3] VRCNet)? It'll be useful for having consistent comparison results.

Thanks~

Hi, thank you for your great work! I'm also recently trying to render some partial point clouds from mesh in my own dataset. I'm using pyrender but it turns out to be super slow. I just wonder if it's possible for you to provide the rendering script? Or may I ask how did you render those partial point clouds (e.g. what software/library did you use) and how long does it take? Thanks!

“We rank each method mainly based on the performance, and you will receive the final ranking list. ” @paul007pl, When would we receive the final ranking list?

I have sent the email of Pravite test to [email protected] and [email protected] yesterday, but I didn't recive a reply. 1. Is it normal to not receive the email about reply? 2. Is there any place to confirm that our documents were successfully received?

Dear Authors,

I noticed that you train the model for completion "using the Adam optimizer with initial learning rate 1e−4 (decayed by 0.7 every 40 epochs) and batch size 32 by NVIDIA TITAN Xp GPU" mentioned in your paper. Did you mean that you used a single GPU for training or more GPUs?

Thanks for your work and look forward to your favourable reply.

While trying to run-> pip install git+https://github.com/open-mmlab/mmdetection.git in the mm3d_pn2 directory, I am getting the following error, how can i fix this?

$ pip install git+https://github.com/open-mmlab/mmdetection.git Collecting git+https://github.com/open-mmlab/mmdetection.git Cloning https://github.com/open-mmlab/mmdetection.git to c:\users\sshre34\appdata\local\temp\pip-req-build-gmxamwua Running command git clone --filter=blob:none --quiet https://github.com/open-mmlab/mmdetection.git 'C:\Users\sshre34\AppData\Local\Temp\pip-req-build-gmxamwua' Resolved https://github.com/open-mmlab/mmdetection.git to commit 94afcdc6aa1e2c849c9d8334ec7730bb52b17922 Preparing metadata (setup.py) ... done Requirement already satisfied: matplotlib in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from mmdet==2.27.0) (3.5.3) Requirement already satisfied: numpy in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from mmdet==2.27.0) (1.21.5) Collecting pycocotools Using cached pycocotools-2.0.6.tar.gz (24 kB) Installing build dependencies ... done Getting requirements to build wheel ... done Preparing metadata (pyproject.toml) ... done Requirement already satisfied: scipy in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from mmdet==2.27.0) (1.7.3) Requirement already satisfied: six in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from mmdet==2.27.0) (1.16.0) Collecting terminaltables Using cached terminaltables-3.1.10-py2.py3-none-any.whl (15 kB) Requirement already satisfied: python-dateutil>=2.7 in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from matplotlib->mmdet==2.27.0) (2.8.2) Requirement already satisfied: pillow>=6.2.0 in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from matplotlib->mmdet==2.27.0) (9.3.0) Requirement already satisfied: fonttools>=4.22.0 in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from matplotlib->mmdet==2.27.0) (4.38.0) Requirement already satisfied: packaging>=20.0 in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from matplotlib->mmdet==2.27.0) (22.0) Requirement already satisfied: kiwisolver>=1.0.1 in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from matplotlib->mmdet==2.27.0) (1.4.4) Requirement already satisfied: cycler>=0.10 in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from matplotlib->mmdet==2.27.0) (0.11.0) Requirement already satisfied: pyparsing>=2.2.1 in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from matplotlib->mmdet==2.27.0) (3.0.9) Requirement already satisfied: typing-extensions in c:\users\sshre34\anaconda3\envs\mvp\lib\site-packages (from kiwisolver>=1.0.1->matplotlib->mmdet==2.27.0) (4.4.0) Building wheels for collected packages: mmdet, pycocotools Building wheel for mmdet (setup.py) ... done Created wheel for mmdet: filename=mmdet-2.27.0-py3-none-any.whl size=1468682 sha256=ee7389458986d8bcfb5c5f486cee48393c239c2200caf362b0b4d31e5a6348ff Stored in directory: C:\Users\sshre34\AppData\Local\Temp\pip-ephem-wheel-cache-iu2x_jfa\wheels\8d\1d\4c\5ba147e9294578f513772158a509161804e08d1b5f62e95705 Building wheel for pycocotools (pyproject.toml) ... error error: subprocess-exited-with-error

× Building wheel for pycocotools (pyproject.toml) did not run successfully. │ exit code: 1 ╰─> [16 lines of output] running bdist_wheel running build running build_py creating build creating build\lib.win-amd64-cpython-37 creating build\lib.win-amd64-cpython-37\pycocotools copying pycocotools\coco.py -> build\lib.win-amd64-cpython-37\pycocotools copying pycocotools\cocoeval.py -> build\lib.win-amd64-cpython-37\pycocotools copying pycocotools\mask.py -> build\lib.win-amd64-cpython-37\pycocotools copying pycocotools_init_.py -> build\lib.win-amd64-cpython-37\pycocotools running build_ext cythoning pycocotools/_mask.pyx to pycocotools_mask.c building 'pycocotools._mask' extension C:\Users\sshre34\AppData\Local\Temp\pip-build-env-02b_xbrm\overlay\Lib\site-packages\Cython\Compiler\Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: C:\Users\sshre34\AppData\Local\Temp\pip-install-3k0didqo\pycocotools_c99855141cd14b4c944e7ec99dcb18aa\pycocotools_mask.pyx tree = Parsing.p_module(s, pxd, full_module_name) error: Microsoft Visual C++ 14.0 or greater is required. Get it with "Microsoft C++ Build Tools": https://visualstudio.microsoft.com/visual-cpp-build-tools/ [end of output]

note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for pycocotools Successfully built mmdet Failed to build pycocotools ERROR: Could not build wheels for pycocotools, which is required to install pyproject.toml-based projects

thank you for your amazing work! i meet a problem and i can not solve it, could you give me some advice? Traceback (most recent call last): File "/raid/MVP_Benchmark/completion/train.py", line 211, in logging.basicConfig(level=logging.INFO, handlers=[logging.FileHandler(os.path.join(log_dir, 'train.log')), File "/home/anaconda3/envs/mvp/lib/python3.7/logging/init.py", line 1087, in init StreamHandler.init(self, self._open()) File "/home/anaconda3/envs/mvp/lib/python3.7/logging/init.py", line 1116, in _open return open(self.baseFilename, self.mode, encoding=self.encoding) FileNotFoundError: [Errno 2] No such file or directory: '/raid/jiangzw/MVP_Benchmark/completion/log/vrcnet_cd_debug_2021-06-18T16:36:20/train.log'

Hy, thanks for your work! During the training of the vrcnet, the following problem occurs:

Especially, the problem occurs whithin the function edge_preserving_sampling, which is called inside the forward of the SA_SKN_Res_encoder:

ds_features, p_idx, pn_idx, ds_points = edge_preserve_sampling(features, points, sample_num, k)

and the feature tensors become full nan tensors. Is there a way to solve this problem? Did I do something wrong during the training?