Comments (8)
Thanks @hengfei-wang for the tip. Thats problem with inplace_abn
come from different cuda version of the system and cuda version of pytorch.
I tried to install pytorch + cudatoolkit==11.3 via conda, but in jupyter i had torch.cuda.is_available() == False
After installing all with pip works like a charm.
My installation process:
conda create -n mvsnerf python=3.8
conda activate mvsnerf
pip install torch==1.10.1+cu113 torchvision==0.11.2+cu113 torchaudio==0.10.1+cu113 -f https://download.pytorch.org/whl/cu113/torch_stable.html
pip install pytorch-lightning==1.3.5 imageio pillow scikit-image opencv-python configargparse lpips kornia warmup_scheduler matplotlib
Installing inplace_abn from source
git clone https://github.com/mapillary/inplace_abn.git
cd inplace_abn
python setup.py install
Dear authors (@apchenstu), please check your repository and notebooks before you publish it, a lot of imports are missed, variables aren't initiated, no dependencies file and full installation requirements. Running just single file become challenging!
from mvsnerf.
besides above, I found that:
- torchmetrics==0.5.0 (otherwise error:
from torchmetrics.utilities.data import get_num_classes as _get_num_classes
ImportError: cannot import name 'get_num_classes' from 'torchmetrics.utilities.data' (/home/anaconda3/envs/mvsnerf/lib/python3.8/site-packages/torchmetrics/utilities/data.py))
)
2. setuptools==59.5.0 (otherwise error:
AttributeError: module 'distutils' has no attribute 'version'
)
3. test-tube==0.7.5 (otherwise error:
Traceback (most recent call last):
File "train_mvs_nerf_finetuning_pl.py", line 302, in <module>
logger = loggers.TestTubeLogger(
File "/home/anaconda3/envs/mvsnerf/lib/python3.8/site-packages/pytorch_lightning/loggers/test_tube.py", line 98, in __init__
raise ImportError(
ImportError: You want to use `test_tube` logger which is not installed yet, install it with `pip install test-tube`.
)
4. pillow==9.5.0 (otherwise error:
AttributeError: module 'PIL.Image' has no attribute 'ANTIALIAS'
)
from mvsnerf.
Well, I used pip install pytorch-lightning==1.3.5 inplace_abn imageio pillow scikit-image opencv-python configargparse lpips kornia==0.5.11 warmup_scheduler
to install packages. But I got 'The CUDA version(11.5) detected is not the one used to compile pytorch(10.5)' error when installing inpalce_abn. Sorry, I can't remember exact output message. I guess pytorch-ligntning installs one version of torch compiled by CUDA 10.2 and inplace_abn detects this version and tries to use it. But it found the local cuda version dismatch with torch version, so it reports the error. I reinstalled pytorch using pip3 install torch==1.10.1+cu113 torchvision==0.11.2+cu113 torchaudio==0.10.1+cu113 -f https://download.pytorch.org/whl/cu113/torch_stable.html
to solve this problem.
Actually, I recommend to use anaconda to install packages in normal cases because conda is very good at managing dependies of different packages. However, there are some packages must be installed using pip in this project. Sad, I have to say using pip troubles me a lot.
from mvsnerf.
I agree. I faced same error... how can i fix it?
from mvsnerf.
@TommeTao hello, did you fix the error? ImportError: /mvsnerf/conda-env-py37/lib/python3.7/site-packages/inplace_abn/_backend.cpython-37m-x86_64-linux-gnu.so: undefined symbol: ZNK2at6Tensor3addEN3c106ScalarES2
from mvsnerf.
@codestella Your error appears, because you compiled some parts of your code (this case inplace_abn) with another pytorch version, than you use later on. This usually happens if you install packages that have a dependency for newer pytorch versions.
What I did was checking which package installs the new pytorch version. In my case this was "kornia" and I needed to specify "pip install kornia==0.5.11".
The dependencies for MVSNeRF that I installed with pip are
pip install pytorch-lightning==1.3.5 inplace_abn imageio pillow scikit-image opencv-python configargparse lpips kornia==0.5.11 warmup_scheduler
I hope that helps you!
from mvsnerf.
Another reminder:
Remember to do 'pip install imageio-ffmpeg' if you want to rendering videos.
Can authors check their code and make a good guideline? :-(
from mvsnerf.
My pip's environment.yml and requirements.txt for you to try and use:
environment.yml
name: mvsnerf
channels:
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/pro/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
- pytorch
- conda-forge
- defaults
dependencies:
- _libgcc_mutex=0.1=main
- _openmp_mutex=5.1=1_gnu
- ca-certificates=2023.05.30=h06a4308_0
- ld_impl_linux-64=2.38=h1181459_1
- libffi=3.4.4=h6a678d5_0
- libgcc-ng=11.2.0=h1234567_1
- libgomp=11.2.0=h1234567_1
- libstdcxx-ng=11.2.0=h1234567_1
- ncurses=6.4=h6a678d5_0
- openssl=3.0.9=h7f8727e_0
- pip=23.1.2=py38h06a4308_0
- python=3.8.17=h955ad1f_0
- readline=8.2=h5eee18b_0
- sqlite=3.41.2=h5eee18b_0
- tk=8.6.12=h1ccaba5_0
- wheel=0.38.4=py38h06a4308_0
- xz=5.4.2=h5eee18b_0
- zlib=1.2.13=h5eee18b_0
- pip:
- absl-py==1.4.0
- aiohttp==3.8.4
- aiosignal==1.3.1
- async-timeout==4.0.2
- attrs==23.1.0
- cachetools==5.3.1
- certifi==2023.5.7
- charset-normalizer==3.2.0
- configargparse==1.5.5
- contourpy==1.1.0
- cycler==0.11.0
- fonttools==4.41.0
- frozenlist==1.4.0
- fsspec==2023.6.0
- future==0.18.3
- google-auth==2.22.0
- google-auth-oauthlib==1.0.0
- grpcio==1.56.0
- idna==3.4
- imageio==2.31.1
- importlib-metadata==6.8.0
- importlib-resources==6.0.0
- inplace-abn==1.1.1.dev7+gd7dd3e1
- kiwisolver==1.4.4
- kornia==0.6.12
- lazy-loader==0.3
- lightning-utilities==0.9.0
- lpips==0.1.4
- markdown==3.4.3
- markupsafe==2.1.3
- matplotlib==3.7.2
- multidict==6.0.4
- networkx==3.1
- numpy==1.24.4
- oauthlib==3.2.2
- opencv-python==4.8.0.74
- pandas==2.0.3
- pillow==9.5.0
- protobuf==4.23.4
- pyasn1==0.5.0
- pyasn1-modules==0.3.0
- pydeprecate==0.3.0
- pyparsing==3.0.9
- python-dateutil==2.8.2
- pytorch-lightning==1.3.5
- pytz==2023.3
- pywavelets==1.4.1
- pyyaml==5.4.1
- requests==2.31.0
- requests-oauthlib==1.3.1
- rsa==4.9
- scikit-image==0.21.0
- scipy==1.10.1
- setuptools==59.5.0
- six==1.16.0
- tensorboard==2.13.0
- tensorboard-data-server==0.7.1
- test-tube==0.7.5
- tifffile==2023.7.10
- torch==1.10.1+cu113
- torchaudio==0.10.1+cu113
- torchmetrics==0.5.0
- torchvision==0.11.2+cu113
- tqdm==4.65.0
- tzdata==2023.3
- urllib3==1.26.16
- warmup-scheduler==0.3
- werkzeug==2.3.6
- yarl==1.9.2
- zipp==3.16.2
requirements.txt
absl-py==1.4.0
aiohttp==3.8.4
aiosignal==1.3.1
async-timeout==4.0.2
attrs==23.1.0
black==23.1.0
cachetools==5.3.1
certifi==2023.5.7
charset-normalizer==3.2.0
click==8.1.3
ConfigArgParse==1.5.5
contourpy==1.1.0
cycler==0.11.0
fonttools==4.41.0
frozenlist==1.4.0
fsspec==2023.6.0
future==0.18.3
google-auth==2.22.0
google-auth-oauthlib==1.0.0
grpcio==1.56.0
idna==3.4
imageio==2.31.1
importlib-metadata==6.8.0
importlib-resources==6.0.0
inplace-abn==1.1.1.dev7+gd7dd3e1
kiwisolver==1.4.4
kornia==0.6.12
lazy_loader==0.3
lightning-utilities==0.9.0
lpips==0.1.4
Markdown==3.4.3
MarkupSafe==2.1.3
matplotlib==3.7.2
multidict==6.0.4
mypy-extensions==1.0.0
networkx==3.1
numpy==1.24.4
oauthlib==3.2.2
opencv-python==4.8.0.74
packaging==23.0
pandas==2.0.3
pathspec==0.11.1
Pillow==9.5.0
platformdirs==3.1.1
protobuf==4.23.4
pyasn1==0.5.0
pyasn1-modules==0.3.0
pyDeprecate==0.3.0
pyparsing==3.0.9
python-dateutil==2.8.2
pytorch-lightning==1.3.5
pytz==2023.3
PyWavelets==1.4.1
PyYAML==5.4.1
requests==2.31.0
requests-oauthlib==1.3.1
rsa==4.9
scikit-image==0.21.0
scipy==1.10.1
six==1.16.0
tensorboard==2.13.0
tensorboard-data-server==0.7.1
test-tube==0.7.5
tifffile==2023.7.10
tomli==2.0.1
torch==1.10.1+cu113
torchaudio==0.10.1+cu113
torchmetrics==0.5.0
torchvision==0.11.2+cu113
tqdm==4.65.0
typing_extensions==4.5.0
tzdata==2023.3
urllib3==1.26.16
warmup-scheduler==0.3
Werkzeug==2.3.6
yarl==1.9.2
zipp==3.16.2
and for the inplace_abn pacakge:
Installing inplace_abn from source
git clone https://github.com/mapillary/inplace_abn.git
cd inplace_abn
python setup.py install
from mvsnerf.
Related Issues (20)
- can't run training command example HOT 3
- why are you not using dir_embedder in the code but writing in the paper ?
- Some problems on rendering dtu HOT 1
- is the ckpt in the directory of ckpts the best of the result HOT 1
- Running on Google Colab and a CKPT bug
- The relationship between DTU training data and Depth_raw HOT 2
- RuntimeError: Function 'PowBackward0' returned nan values in its 0th output.
- ValueError: The provided lr scheduler "<torch.optim.lr_scheduler.MultiStepLR object at 0x000001FE8239EAF0>" is invalid HOT 3
- TypeError: cannot unpack non-iterable NoneType object HOT 1
- Test the DTU dataset with IBRNet
- How many views are used during finetuning?
- Each run reports a different error: No such file or directory: './dataset/dtu/Rectified/scan1_train/rect_011_3_r5000.png' HOT 1
- How can I get the new depth map
- win_size exceeds image extent. Either ensure that your images are at least 7x7;
- Pre-trained checkpoint
- When train llff, keyerror 'w2cs' occurs in validation step
- Depth estimation
- run training example
- Depth resolution
- how to train non per scene optimization mvsnerf?
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.
from mvsnerf.