fwilliams / deep-geometric-prior Goto Github PK
View Code? Open in Web Editor NEWThe reference implementaiton for the paper "Deep Geometric Prior for Surface Reconstruction"
License: MIT License
The reference implementaiton for the paper "Deep Geometric Prior for Surface Reconstruction"
License: MIT License
Hi, I've noticed your utils.py seems to be outdated.
Since I tried to run the reconstruct_single_patch.py and found the module pcu has no methods called read_off/read_ply...
I suppose it's about you having updated pcu module but not utils.py?
Hi, thanks for your work.
I found the 'make_patch' is very time-consuming for the SRB benchmark and for some reason, I do not have enough time to run it....
Could you share the results under SRB?
Greatly appreciate it, I will never forget your favor.
Hi, following the instructions I get this error in console:
(base) PS C:\Users\G\Desktop\PYTHON_graphs> cd C:\Users\G\repositories\deep-geometric-prior
(base) PS C:\Users\G\repositories\deep-geometric-prior> conda env create -f environment.yml
Collecting package metadata (repodata.json): done
Solving environment: failed
ResolvePackageNotFound:
- qt==5.6.3=h8bf5577_3
- libffi==3.2.1=hd88cf55_4
- libxcb==1.13=h14c3975_1002
- ptyprocess==0.6.0=py37_0
- xorg-xproto==7.0.31=h14c3975_1007
- kiwisolver==1.1.0=py37hc9558a2_0
- pytorch==1.1.0=py3.7_cuda9.0.176_cudnn7.5.1_0
- pillow==6.2.1=py37h34e0f95_0
- libstdcxx-ng==9.1.0=hdf63c60_0
- sqlite==3.30.1=h7b6447c_0
- torchvision==0.3.0=py37_cu9.0.176_1
- vtk==8.2.0=py37ha8e561a_201
- xorg-libx11==1.6.9=h516909a_0
- libnetcdf==4.6.2=hbdf4f91_1001
- openssl==1.1.1d=h516909a_0
- xorg-libxt==1.2.0=h516909a_0
- mayavi==4.7.1=py37h7eb8c7e_2
- libssh2==1.8.2=h22169c7_2
- xorg-libsm==1.2.3=h84519dc_1000
- icu==58.2=hf484d3e_1000
- traits==5.2.0=py37h516909a_0
- zstd==1.3.7=h0b5b093_0
- glib==2.58.3=py37h6f030ca_1002
- zlib==1.2.11=h7b6447c_3
- tornado==6.0.3=py37h516909a_0
- numpy==1.17.3=py37hd14ec0e_0
- numpy-base==1.17.3=py37hde5b4d6_0
- bzip2==1.0.8=h516909a_1
- matplotlib==2.2.2=py37hb69df0a_2
- readline==7.0=h7b6447c_5
- tk==8.6.8=hbc83047_0
- mkl_random==1.1.0=py37hd6b4f25_0
- cudatoolkit==9.0=h13b8566_0
- mkl==2019.4=243
- dbus==1.13.6=he372182_0
- pcre==8.43=he1b5a44_0
- fontconfig==2.13.1=he4413a7_1000
- libgfortran-ng==7.3.0=hdf63c60_0
- pyqt==5.6.0=py37h13b7fb3_1008
- xz==5.2.4=h14c3975_4
- gst-plugins-base==1.14.5=h0935bb2_0
- hdf5==1.10.4=nompi_h3c11f04_1106
- scipy==1.3.1=py37h7c811a0_0
- jpeg==9c=h14c3975_1001
- python==3.7.5=h0371630_0
- xorg-libice==1.0.10=h516909a_0
- xorg-libxau==1.0.9=h14c3975_0
- expat==2.2.5=he1b5a44_1004
- libtiff==4.1.0=h2733197_0
- freetype==2.9.1=h8a8886c_1
- jsoncpp==1.8.4=hc9558a2_1002
- libpng==1.6.37=hbc83047_0
- pthread-stubs==0.4=h14c3975_1001
- sip==4.18.1=py37hf484d3e_1000
- hdf4==4.2.13=hf30be14_1003
- lz4-c==1.8.3=he1b5a44_1001
- libgcc-ng==9.1.0=hdf63c60_0
- libiconv==1.15=h516909a_1005
- mkl-service==2.3.0=py37he904b0f_0
- ncurses==6.1=he6710b0_1
- ninja==1.9.0=py37hfd86e86_0
- xorg-kbproto==1.0.7=h14c3975_1002
- point_cloud_utils==0.12.0=py37h9de70de_0
- libcurl==7.62.0=h20c2e04_0
- cffi==1.13.1=py37h2e261b9_0
- libuuid==2.32.1=h14c3975_1000
- libxml2==2.9.9=h13577e0_2
- gettext==0.19.8.1=hc5be6a0_1002
- gstreamer==1.14.5=h36ae1b5_0
- tbb==2019.9=hc9558a2_0
- xorg-libxdmcp==1.1.3=h516909a_0
- intel-openmp==2019.4=243
- libedit==3.1.20181209=hc058e9b_0
- curl==7.62.0=hbc83047_0
- mkl_fft==1.0.15=py37ha843d7b_0
I have updated conda.
win 10 64 bit
Python 3.8.6 | packaged by conda-forge | (default, Oct 7 2020, 18:22:52) [MSC v.1916 64 bit (AMD64)] on win32
The file plot_reconstruction.py seems to be missing.
Hi,
thank you for making this great project available!
May I ask where you found the ground truth surfaces of the benchmark (Gargoyle, Quasimoto, Daratech etc.)?
I would like to try out samplings with different scanner options applied to the models, but for this I need the ground truth surfaces.
It seems like they are not directly provided in the benchmark data.
Hello I am
Running this code on a single gpu GTX 1080 Ti and getting this error:
RuntimeError: CUDA out of memory. Tried to allocate 2.00 MiB (GPU 0; 10.92 GiB total capacity; 10.36 GiB already allocated; 3.00 MiB free; 14.60 MiB cached)
The traceback
File "reconstruct_surface.py", line 495, in <module>
main()
File "reconstruct_surface.py", line 416, in main
sum_loss.backward()
File "/home/azad/.conda/envs/deep-geometric-prior/lib/python3.7/site-packages/torch/tensor.py", line 107, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph)
File "/home/azad/.conda/envs/deep-geometric-prior/lib/python3.7/site-packages/torch/autograd/__init__.py", line 93, in backward
allow_unreachable=True) # allow_unreachable flag
Any thoughts on how to make it work? Should I downsample the point cloud? I used the lord_quas.ply which has the least size of all the others.
Hi there,
Thank you for release the codes for this amazing work! I wonder why the export_point_cloud.py is empty, will you update it later or ?
Thanks,
Ryan
Thanks for your impressive CVPR2019 work and kindly sharing your code. I wonder if you could provide some test data and a quick demo instruction so that we can see the result in your paper? Thanks again!
Hi, thanks for sharing this codebase!
I was wondering if you could share the code for plotting? I was hoping to try and fix it, assuming it is still broken.
how can i implment this work on Colab?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.