czvvd / svdformer_pointsea Goto Github PK

View Code? Open in Web Editor NEW

44.0 44.0 3.0 4.11 MB

[ICCV 2023] SVDFormer: Complementing Point Cloud via Self-view Augmentation and Self-structure Dual-generator

License: MIT License

Python 80.40% Cuda 13.29% C++ 5.39% Shell 0.31% C 0.61%

svdformer_pointsea's People

Contributors

Stargazers

Watchers

Forkers

one-of-student mmonsterr chenhonghua th-huang

svdformer_pointsea's Issues

Cannot reproduce the result directly

Thanks for your amazing work!

If I only change the some basic setting in the config (like dataset path), and run main_pcn directly, the test result doesn't go as expected. It seems that the training process doesn't work.

Here is the log output

2023-10-16 00:57:28,897 - INFO - 'epoch: ', 1, 'optimizer: ', 0.0
2023-10-16 01:37:58,150 - INFO - [Epoch 1/400] EpochTime = 2429.254 (s) Losses = ['47.3596', '22.5916', '18.0766']
2023-10-16 01:38:52,107 - INFO - ============================ TEST RESULTS ============================
2023-10-16 01:38:52,107 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 01:38:52,108 - INFO - 02691156 13 11.3990 0.7054 0.6294
2023-10-16 01:38:52,108 - INFO - 02933112 12 20.8394 0.7446 0.3774
2023-10-16 01:38:52,108 - INFO - 02958343 13 14.9069 0.7059 0.4695
2023-10-16 01:38:52,109 - INFO - 03001627 12 18.8508 0.7260 0.4516
2023-10-16 01:38:52,109 - INFO - 03636649 13 17.8343 0.7393 0.4878
2023-10-16 01:38:52,109 - INFO - 04256520 12 21.6226 0.7700 0.3663
2023-10-16 01:38:52,119 - INFO - 04379243 13 15.1328 0.6857 0.5245
2023-10-16 01:38:52,119 - INFO - 04530566 12 13.5874 0.7076 0.5510
2023-10-16 01:38:52,120 - INFO - Overall 16.6935 0.7225 0.4840

2023-10-16 01:38:52,120 - INFO - Epoch 1 16.6935 0.7225 0.4840

2023-10-16 01:38:53,844 - INFO - Saved checkpoint to SVDFormer-main/experiment/ckpt\ckpt-best.pth ...
2023-10-16 01:38:53,844 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 01:38:53,845 - INFO - 'epoch: ', 2, 'optimizer: ', 0.001
2023-10-16 02:18:53,104 - INFO - [Epoch 2/400] EpochTime = 2399.259 (s) Losses = ['37.4754', '19.4601', '102.9291']
2023-10-16 02:19:45,864 - INFO - ============================ TEST RESULTS ============================
2023-10-16 02:19:45,864 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 02:19:45,864 - INFO - 02691156 13 11.5185 0.9148 0.6104
2023-10-16 02:19:45,864 - INFO - 02933112 12 22.4493 0.9322 0.2161
2023-10-16 02:19:45,864 - INFO - 02958343 13 18.0634 0.9226 0.2638
2023-10-16 02:19:45,864 - INFO - 03001627 12 20.2612 0.9288 0.3075
2023-10-16 02:19:45,864 - INFO - 03636649 13 18.2606 0.9323 0.4070
2023-10-16 02:19:45,864 - INFO - 04256520 12 22.6796 0.9337 0.2306
2023-10-16 02:19:45,864 - INFO - 04379243 13 16.2068 0.9182 0.3807
2023-10-16 02:19:45,864 - INFO - 04530566 12 14.9223 0.9204 0.4352
2023-10-16 02:19:45,864 - INFO - Overall 17.9639 0.9252 0.3588

2023-10-16 02:19:45,864 - INFO - Epoch 2 17.9639 0.9252 0.3588

2023-10-16 02:19:45,864 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 02:19:45,864 - INFO - 'epoch: ', 3, 'optimizer: ', 0.001
2023-10-16 02:59:26,815 - INFO - [Epoch 3/400] EpochTime = 2380.950 (s) Losses = ['61.0343', '130.7800', '128.8345']
2023-10-16 03:00:19,536 - INFO - ============================ TEST RESULTS ============================
2023-10-16 03:00:19,536 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 03:00:19,536 - INFO - 02691156 13 18.2846 0.9746 0.4420
2023-10-16 03:00:19,536 - INFO - 02933112 12 31.3285 0.9778 0.1221
2023-10-16 03:00:19,536 - INFO - 02958343 13 27.4990 0.9778 0.1485
2023-10-16 03:00:19,536 - INFO - 03001627 12 25.8402 0.9779 0.2059
2023-10-16 03:00:19,536 - INFO - 03636649 13 21.2641 0.9762 0.4088
2023-10-16 03:00:19,536 - INFO - 04256520 12 32.3640 0.9780 0.1478
2023-10-16 03:00:19,536 - INFO - 04379243 13 22.2649 0.9776 0.2051
2023-10-16 03:00:19,536 - INFO - 04530566 12 19.3846 0.9752 0.3185
2023-10-16 03:00:19,536 - INFO - Overall 24.6807 0.9769 0.2519

2023-10-16 03:00:19,536 - INFO - Epoch 3 24.6807 0.9769 0.2519

2023-10-16 03:00:19,536 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 03:00:19,536 - INFO - 'epoch: ', 4, 'optimizer: ', 0.001
2023-10-16 03:39:52,293 - INFO - [Epoch 4/400] EpochTime = 2372.757 (s) Losses = ['52.4664', '25.0615', '23.0664']
2023-10-16 03:40:45,025 - INFO - ============================ TEST RESULTS ============================
2023-10-16 03:40:45,025 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 03:40:45,025 - INFO - 02691156 13 16.3487 0.9754 0.4308
2023-10-16 03:40:45,025 - INFO - 02933112 12 29.5060 0.9789 0.1103
2023-10-16 03:40:45,025 - INFO - 02958343 13 22.8731 0.9781 0.1376
2023-10-16 03:40:45,025 - INFO - 03001627 12 24.0258 0.9792 0.1787
2023-10-16 03:40:45,025 - INFO - 03636649 13 22.2710 0.9788 0.3626
2023-10-16 03:40:45,025 - INFO - 04256520 12 27.1255 0.9786 0.1333
2023-10-16 03:40:45,025 - INFO - 04379243 13 22.3254 0.9788 0.1864
2023-10-16 03:40:45,025 - INFO - 04530566 12 18.4356 0.9766 0.2924
2023-10-16 03:40:45,025 - INFO - Overall 22.7875 0.9781 0.2310

2023-10-16 03:40:45,025 - INFO - Epoch 4 22.7875 0.9781 0.2310

2023-10-16 03:40:45,025 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 03:40:45,025 - INFO - 'epoch: ', 5, 'optimizer: ', 0.001
2023-10-16 04:20:17,369 - INFO - [Epoch 5/400] EpochTime = 2372.343 (s) Losses = ['47.6701', '24.3645', '22.4079']
2023-10-16 04:21:10,066 - INFO - ============================ TEST RESULTS ============================
2023-10-16 04:21:10,066 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 04:21:10,066 - INFO - 02691156 13 15.6621 0.9750 0.4384
2023-10-16 04:21:10,066 - INFO - 02933112 12 29.6768 0.9786 0.1132
2023-10-16 04:21:10,066 - INFO - 02958343 13 22.7800 0.9779 0.1396
2023-10-16 04:21:10,066 - INFO - 03001627 12 23.3449 0.9788 0.1872
2023-10-16 04:21:10,066 - INFO - 03636649 13 21.1941 0.9780 0.3753
2023-10-16 04:21:10,066 - INFO - 04256520 12 27.7056 0.9784 0.1371
2023-10-16 04:21:10,066 - INFO - 04379243 13 20.6201 0.9777 0.1997
2023-10-16 04:21:10,066 - INFO - 04530566 12 17.4098 0.9762 0.2985
2023-10-16 04:21:10,066 - INFO - Overall 22.2098 0.9776 0.2382

2023-10-16 04:21:10,066 - INFO - Epoch 5 22.2098 0.9776 0.2382

2023-10-16 04:21:10,066 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 04:21:10,066 - INFO - 'epoch: ', 6, 'optimizer: ', 0.001
2023-10-16 05:00:43,427 - INFO - [Epoch 6/400] EpochTime = 2373.361 (s) Losses = ['48.0799', '24.5652', '22.6182']
2023-10-16 05:01:36,117 - INFO - ============================ TEST RESULTS ============================
2023-10-16 05:01:36,117 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 05:01:36,117 - INFO - 02691156 13 15.9628 0.9753 0.4295
2023-10-16 05:01:36,117 - INFO - 02933112 12 29.7972 0.9794 0.1073
2023-10-16 05:01:36,117 - INFO - 02958343 13 22.4433 0.9783 0.1341
2023-10-16 05:01:36,117 - INFO - 03001627 12 24.8605 0.9795 0.1770
2023-10-16 05:01:36,117 - INFO - 03636649 13 22.1154 0.9789 0.3588
2023-10-16 05:01:36,117 - INFO - 04256520 12 27.2313 0.9787 0.1317
2023-10-16 05:01:36,117 - INFO - 04379243 13 21.7234 0.9787 0.1848
2023-10-16 05:01:36,117 - INFO - 04530566 12 17.6590 0.9768 0.2859
2023-10-16 05:01:36,117 - INFO - Overall 22.6376 0.9782 0.2282

2023-10-16 05:01:36,117 - INFO - Epoch 6 22.6376 0.9782 0.2282

2023-10-16 05:01:36,117 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 05:01:36,117 - INFO - 'epoch: ', 7, 'optimizer: ', 0.001
2023-10-16 05:41:14,628 - INFO - [Epoch 7/400] EpochTime = 2378.511 (s) Losses = ['59.4867', '25.3906', '23.3751']
2023-10-16 05:42:07,531 - INFO - ============================ TEST RESULTS ============================
2023-10-16 05:42:07,531 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 05:42:07,531 - INFO - 02691156 13 19.1921 0.9720 0.4686
2023-10-16 05:42:07,531 - INFO - 02933112 12 45.4983 0.9759 0.1426
2023-10-16 05:42:07,531 - INFO - 02958343 13 39.2203 0.9760 0.1704
2023-10-16 05:42:07,531 - INFO - 03001627 12 27.5350 0.9748 0.2490
2023-10-16 05:42:07,531 - INFO - 03636649 13 21.0602 0.9727 0.4666
2023-10-16 05:42:07,531 - INFO - 04256520 12 44.8739 0.9760 0.1736
2023-10-16 05:42:07,531 - INFO - 04379243 13 21.7614 0.9744 0.2468
2023-10-16 05:42:07,531 - INFO - 04530566 12 23.6168 0.9733 0.3510
2023-10-16 05:42:07,531 - INFO - Overall 30.1433 0.9744 0.2858

2023-10-16 05:42:07,531 - INFO - Epoch 7 30.1433 0.9744 0.2858

2023-10-16 05:42:07,531 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 05:42:07,531 - INFO - 'epoch: ', 8, 'optimizer: ', 0.001
2023-10-16 06:21:44,065 - INFO - [Epoch 8/400] EpochTime = 2376.533 (s) Losses = ['62.6372', '26.2596', '24.1856']
2023-10-16 06:22:36,760 - INFO - ============================ TEST RESULTS ============================
2023-10-16 06:22:36,760 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 06:22:36,760 - INFO - 02691156 13 17.1301 0.9754 0.4338
2023-10-16 06:22:36,760 - INFO - 02933112 12 30.7366 0.9786 0.1168
2023-10-16 06:22:36,760 - INFO - 02958343 13 24.6502 0.9780 0.1437
2023-10-16 06:22:36,760 - INFO - 03001627 12 24.7679 0.9782 0.2011
2023-10-16 06:22:36,760 - INFO - 03636649 13 22.4678 0.9775 0.3907
2023-10-16 06:22:36,760 - INFO - 04256520 12 28.2933 0.9778 0.1457
2023-10-16 06:22:36,760 - INFO - 04379243 13 21.2204 0.9775 0.2030
2023-10-16 06:22:36,760 - INFO - 04530566 12 18.0671 0.9759 0.3071
2023-10-16 06:22:36,760 - INFO - Overall 23.3347 0.9773 0.2447

2023-10-16 06:22:36,760 - INFO - Epoch 8 23.3347 0.9773 0.2447

2023-10-16 06:22:36,760 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 06:22:36,760 - INFO - 'epoch: ', 9, 'optimizer: ', 0.001
2023-10-16 07:02:13,236 - INFO - [Epoch 9/400] EpochTime = 2376.476 (s) Losses = ['50.3192', '24.6044', '22.6177']
2023-10-16 07:03:05,981 - INFO - ============================ TEST RESULTS ============================
2023-10-16 07:03:05,981 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 07:03:05,981 - INFO - 02691156 13 17.2960 0.9755 0.4323
2023-10-16 07:03:05,981 - INFO - 02933112 12 30.2031 0.9782 0.1174
2023-10-16 07:03:05,981 - INFO - 02958343 13 24.6921 0.9783 0.1409
2023-10-16 07:03:05,981 - INFO - 03001627 12 24.2498 0.9785 0.1930
2023-10-16 07:03:05,981 - INFO - 03636649 13 20.9219 0.9775 0.3840
2023-10-16 07:03:05,981 - INFO - 04256520 12 28.4580 0.9782 0.1404
2023-10-16 07:03:05,981 - INFO - 04379243 13 21.0785 0.9778 0.1979
2023-10-16 07:03:05,981 - INFO - 04530566 12 18.1918 0.9763 0.2991
2023-10-16 07:03:05,981 - INFO - Overall 23.0508 0.9775 0.2402

2023-10-16 07:03:05,981 - INFO - Epoch 9 23.0508 0.9775 0.2402

2023-10-16 07:03:05,981 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 07:03:05,981 - INFO - 'epoch: ', 10, 'optimizer: ', 0.001
2023-10-16 07:42:39,156 - INFO - [Epoch 10/400] EpochTime = 2373.176 (s) Losses = ['49.9786', '24.7791', '22.8088']
2023-10-16 07:43:31,821 - INFO - ============================ TEST RESULTS ============================
2023-10-16 07:43:31,821 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 07:43:31,821 - INFO - 02691156 13 16.1790 0.9751 0.4319
2023-10-16 07:43:31,821 - INFO - 02933112 12 29.5857 0.9783 0.1150
2023-10-16 07:43:31,821 - INFO - 02958343 13 23.4589 0.9779 0.1413
2023-10-16 07:43:31,821 - INFO - 03001627 12 24.6318 0.9789 0.1869
2023-10-16 07:43:31,821 - INFO - 03636649 13 21.5646 0.9778 0.3769
2023-10-16 07:43:31,821 - INFO - 04256520 12 28.1176 0.9785 0.1371
2023-10-16 07:43:31,821 - INFO - 04379243 13 21.8681 0.9781 0.1885
2023-10-16 07:43:31,821 - INFO - 04530566 12 17.9020 0.9763 0.2955
2023-10-16 07:43:31,821 - INFO - Overall 22.8276 0.9776 0.2361

2023-10-16 07:43:31,821 - INFO - Epoch 10 22.8276 0.9776 0.2361

2023-10-16 07:43:31,821 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 07:43:31,821 - INFO - 'epoch: ', 11, 'optimizer: ', 0.001
2023-10-16 08:23:06,061 - INFO - [Epoch 11/400] EpochTime = 2374.239 (s) Losses = ['48.8284', '24.2318', '22.2491']
2023-10-16 08:23:58,873 - INFO - ============================ TEST RESULTS ============================
2023-10-16 08:23:58,873 - INFO - Taxonomy #Sample CD DCD F1

2023-10-16 08:23:58,873 - INFO - 02691156 13 16.9094 0.9752 0.4345
2023-10-16 08:23:58,873 - INFO - 02933112 12 29.8556 0.9785 0.1145
2023-10-16 08:23:58,873 - INFO - 02958343 13 24.1074 0.9783 0.1389
2023-10-16 08:23:58,873 - INFO - 03001627 12 24.4389 0.9792 0.1858
2023-10-16 08:23:58,873 - INFO - 03636649 13 22.1557 0.9785 0.3676
2023-10-16 08:23:58,873 - INFO - 04256520 12 28.3786 0.9787 0.1378
2023-10-16 08:23:58,873 - INFO - 04379243 13 24.9475 0.9799 0.1859
2023-10-16 08:23:58,873 - INFO - 04530566 12 18.6872 0.9765 0.2950
2023-10-16 08:23:58,873 - INFO - Overall 23.6188 0.9781 0.2345

2023-10-16 08:23:58,873 - INFO - Epoch 11 23.6188 0.9781 0.2345

2023-10-16 08:23:58,873 - INFO - Best Performance: Epoch 1 -- CD 16.6935
2023-10-16 08:23:58,873 - INFO - 'epoch: ', 12, 'optimizer: ', 0.001

ModuleNotFoundError: No module named 'torchvision.models.utils'

Hi,
I am trying to run training using PCN dataset, I have got this following error.

from torchvision.models.utils import load_state_dict_from_url
ModuleNotFoundError: No module named 'torchvision.models.utils'

not being able to install pointNet++

Hello,

I followed the instructions for creating a conda environment with CUFDA 11.6 python version 3.9 and pytorch version 1.13 but when I execute this command at the beginning : cd pointnet2_ops_lib
python setup.py install

It gives me the following error : File "/home/safebot/anaconda3/envs/svdformer/lib/python3.9/site-packages/setuptools/command/build_ext.py", line 84, in run
_build_ext.run(self)
File "/home/safebot/anaconda3/envs/svdformer/lib/python3.9/site-packages/setuptools/_distutils/command/build_ext.py", line 345, in run
self.build_extensions()
File "/home/safebot/anaconda3/envs/svdformer/lib/python3.9/site-packages/torch/utils/cpp_extension.py", line 499, in build_extensions
_check_cuda_version(compiler_name, compiler_version)
File "/home/safebot/anaconda3/envs/svdformer/lib/python3.9/site-packages/torch/utils/cpp_extension.py", line 382, in _check_cuda_version
torch_cuda_version = packaging.version.parse(torch.version.cuda)
File "/home/safebot/anaconda3/envs/svdformer/lib/python3.9/site-packages/pkg_resources/_vendor/packaging/version.py", line 52, in parse
return Version(version)
File "/home/safebot/anaconda3/envs/svdformer/lib/python3.9/site-packages/pkg_resources/_vendor/packaging/version.py", line 196, in init
match = self._regex.search(version)
TypeError: expected string or bytes-like object
could you please help me and tell me where do you think the issue is? I think it is version.py but that is just one string variable when I checked inside .
kind regards
Nima

Training Time Questiones

Hi authors, thanks for your great work,

I try to reproduce your work and when I tried to do train it, I found it is very slow to run one epoch, so, I want to know what devices do you use and how long time for you to train one epoch and the whole experiment. Thanks very much

Use our dataset

May I ask how we should use our own dataset and perform complementary experiments with this amazing pre-trained weights of yours? We would like to get the result of inputting only the mutilated point cloud and outputting the complete point cloud

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.