Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Bug] Unable to build opencv container with the default opencv tag 4.8.1 #591

Open
hrishi121 opened this issue Aug 16, 2024 · 2 comments
Open

Comments

@hrishi121
Copy link

hrishi121 commented Aug 16, 2024

Hello,

I have been trying to build the NanoLLM container for following system config:

Jetson Orin NX
-- L4T_VERSION=36.3.0
-- JETPACK_VERSION=6.0
-- CUDA_VERSION=12.4
-- PYTHON_VERSION=3.10
-- LSB_RELEASE=22.04 (jammy)

One of the dependencies for the NanoLLM container is the opencv container. When building the opencv container, I noticed that install.sh would fail and it would proceed to build the opencv-python repo using the build.sh script. However, compiling opencv-python from scratch with the build.sh would keep failing as well (Unfortunately, I forgot to capture the build errors)

I commented out the following lines in build.sh, and was able to build the opencv container

#git checkout --recurse-submodules ${OPENCV_VERSION}
#cat modules/core/include/opencv2/core/version.hpp

#cd ../opencv_contrib
#git checkout --recurse-submodules ${OPENCV_VERSION}

#cd ../opencv_extra
#git checkout --recurse-submodules ${OPENCV_VERSION}

Commenting out these lines disables the build from using the 4.8.1 submodules for opencv, opencv_contrib, opencv_extras and rather use the new default 4.10.0 submodules in the 4.x branch

I am not totally sure why the build was failing earlier, but my guesses are:

  • Incompatibility of CUDA 12.4 with the 4.8.1 submodules
  • [OR] Incompatibility of the latest opencv-python branch 4.x with 4.8.1 submodules

Just wanted to give a heads-up to anyone who is building any container that has the opencv as a dependency and using CUDA 12.4

@hrishi121
Copy link
Author

@dusty-nv Is there a specific reason for using 4.8.1 as the default opencv version in config.py ? Do you think it would be a good idea to update it to 4.10.0 ?

@johnnynunez
Copy link
Contributor

@dusty-nv Is there a specific reason for using 4.8.1 as the default opencv version in config.py ? Do you think it would be a good idea to update it to 4.10.0 ?

FYI I'm using opencv 4.10 with cuda 12.6 cudnn last version without problems

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants