On the tutorial https://github.com/facebookresearch/pytorch3d/blob/master/docs/tutorials/render_colored_points.ipynb, now that the name has changed from "visualization' to 'vis' there is a _ModuleNotFoundError: No module named 'pytorch3d.vis'_
<import os
import torch
import torch.nn.functional as F
import matplotlib.pyplot as plt
from skimage.io import imread
import numpy as np
from pytorch3d.structures import Pointclouds
from pytorch3d.vis import AxisArgs, plot_pointclouds
from pytorch3d.renderer import (
look_at_view_transform,
FoVOrthographicCameras,
PointsRasterizationSettings,
PointsRenderer,
PointsRasterizer,
AlphaCompositor,
NormWeightedCompositor
)>
git diff) or code you wrote<from pytorch3d.visualization import AxisArgs, plot_pointclouds>
<ModuleNotFoundError Traceback (most recent call last)
<ipython-input-3-445b2bac4bbb> in <module>()
10 # Data structures and functions for rendering
11 from pytorch3d.structures import Pointclouds
---> 12 from pytorch3d.visualization import AxisArgs, plot_pointclouds
13 from pytorch3d.renderer import (
14 look_at_view_transform,
ModuleNotFoundError: No module named 'pytorch3d.visualization'
---------------------------------------------------------------------------
NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.
To view examples of installing some common dependencies, click the
"Open Examples" button below.>
Are you running the tutorial after cloning the github repo or on Colab? It depends how you are installing PyTorch3D. The conda packages do not have the vis module.
Thanks @nikhilaravi for the quick reply.
I using from Google Colab,
I run this before for installing pytorch3d, from the stable as reported in the tutorial
!pip install torch torchvision
import sys
import torch
if torch.__version__=='1.6.0+cu101' and sys.platform.startswith('linux'):
!pip install pytorch3d
else:
!pip install 'git+https://github.com/facebookresearch/pytorch3d.git@stable'
As explained in https://github.com/facebookresearch/pytorch3d/issues/385#issuecomment-704485523, we have several experimental features in the master branch but not in the conda package - the stable tag references the commit which was used to build the conda package. If you want to use the code available in master you need to remove the @stable tag at the end of the pip install command.
Thanks @nikhilaravi for the reference, it didn't appear on my initial search unfortunately
I followed your suggestion and I used only this line
!pip install 'git+https://github.com/facebookresearch/pytorch3d.git' instead of the usual
!pip install torch torchvision
import sys
import torch
if torch.__version__=='1.6.0+cu101' and sys.platform.startswith('linux'):
!pip install pytorch3d
else:
!pip install 'git+https://github.com/facebookresearch/pytorch3d.git@stable'
it took around 7min
and then Successfully built pytorch3d
unfortunately I still have the same error
ModuleNotFoundError: No module named 'pytorch3d.vis'
I tried also with .visualization but same error
Looking at 0.2.5, which as of today is the version currently in pypi
https://github.com/facebookresearch/pytorch3d/tree/v0.2.5/pytorch3d
v0.2.5 definately does not contain pytorch3d.vis
building from source using pip also didn't work for me
@DuaneNielsen @albertotono ok I will take a look at this and get back to you!
The change was from pytorch3d.visualization to pytorch3d.vis, not the other way around.
I think there's something strange about pip installing from github in to colab. Several times, I tried
!pip install 'git+https://github.com/facebookresearch/pytorch3d.git'
from pytorch3d.vis import AxisArgs, plot_pointclouds
in a blank colab notebook. It should work, and it usually did, but it did raise your error once, maybe due to a caching issue.
Can you maybe restart and try again?
Thank you all, confirmed that restarting the runtime didn't work , but using a new notebook worked perfectly! I am closing this issue. Thank you so much.
Most helpful comment
@DuaneNielsen @albertotono ok I will take a look at this and get back to you!