Support for CUDA11.x?

[mengxing233-cloud]

Does PyCuVSLAM support running on CUDA 11.x? I want to integrate some segmentation models with PyCuVSLAM. Due to the requirements of the mmsegmentation framework, I need to run it on something like CUDA 11.5.

[zwdoescode]

PyCuVSLAM is the python wrapper for cuVSLAM. Currently cuVSLAM only supports CUDA 12.6.
There is a lot of API changes between CUDA 11.5 and 12.6. So unfortunately PyCuVSLAM doesn't work for CUDA 11.5 right now.

[mengxing233-cloud]

@zwdoescode Thank you for your reply! I have some questions. I have configured CUDA 11.5 in a Conda environment and tried to run projects in that environment. I tested on the EuroC and Kitti datasets, and it seems to work without any issues. This is somewhat puzzling. My environment and result are as follows:
environment ：

Name Version Build Channel

_libgcc_mutex 0.1 main
_openmp_mutex 5.1 1_gnu
argon2-cffi 25.1.0 pypi_0 pypi
argon2-cffi-bindings 25.1.0 pypi_0 pypi
boto3 1.42.9 pypi_0 pypi
botocore 1.42.9 pypi_0 pypi
bzip2 1.0.8 h5eee18b_6
ca-certificates 2025.12.2 h06a4308_0
certifi 2025.11.12 pypi_0 pypi
cffi 2.0.0 pypi_0 pypi
charset-normalizer 3.4.4 pypi_0 pypi
colorama 0.4.6 pypi_0 pypi
cudatoolkit 11.5.2 hbdc67f6_13 conda-forge
cudnn 8.3.2.44 hed8a83a_1 conda-forge
cupy-cuda11x 13.6.0 pypi_0 pypi
cuvslam 0.2.0 pypi_0 pypi
einops 0.8.1 pypi_0 pypi
expat 2.7.3 h7354ed3_4
fastrlock 0.8.3 pypi_0 pypi
filelock 3.20.0 pypi_0 pypi
hf-xet 1.2.0 pypi_0 pypi
huggingface-hub 0.17.3 pypi_0 pypi
idna 3.11 pypi_0 pypi
imageio 2.37.2 pypi_0 pypi
jmespath 1.0.1 pypi_0 pypi
kornia 0.8.2 pypi_0 pypi
kornia-rs 0.1.10 pypi_0 pypi
ld_impl_linux-64 2.44 h153f514_2
libexpat 2.7.3 h7354ed3_4
libffi 3.4.4 h6a678d5_1
libgcc 15.2.0 h767d61c_6 conda-forge
libgcc-ng 15.2.0 h69a702a_6 conda-forge
libgomp 15.2.0 h767d61c_6 conda-forge
libnsl 2.0.0 h5eee18b_0
libstdcxx 15.2.0 h8f9b012_6 conda-forge
libstdcxx-ng 12.2.0 h46fd767_19 conda-forge
libuuid 1.41.5 h5eee18b_0
libxcb 1.17.0 h9b100fa_0
libzlib 1.3.1 hb25bd0a_0
llvmlite 0.46.0 pypi_0 pypi
lxml 6.0.2 pypi_0 pypi
markupsafe 3.0.3 pypi_0 pypi
minio 7.2.20 pypi_0 pypi
ncurses 6.5 h7934f7d_0
numba 0.63.1 pypi_0 pypi
numpy 1.22.0 pypi_0 pypi
opencv-contrib-python 4.12.0.88 pypi_0 pypi
openssl 3.6.0 h26f9b46_0 conda-forge
packaging 25.0 pypi_0 pypi
pandas 2.3.3 pypi_0 pypi
pillow 11.1.0 pypi_0 pypi
pip 25.3 pyhc872135_0
plyfile 1.1.3 pypi_0 pypi
protobuf 6.33.2 pypi_0 pypi
pthread-stubs 0.3 h0ce48e5_1
pyarrow 22.0.0 pypi_0 pypi
pycparser 2.23 pypi_0 pypi
pycryptodome 3.23.0 pypi_0 pypi
pyquaternion 0.9.9 pypi_0 pypi
python 3.10.19 h6fa692b_0
pytransform3d 3.14.4 pypi_0 pypi
pytz 2025.2 pypi_0 pypi
pyyaml 6.0.2 pypi_0 pypi
readline 8.3 hc2a1206_0
regex 2025.11.3 pypi_0 pypi
requests 2.32.5 pypi_0 pypi
rerun-sdk 0.22.1 pypi_0 pypi
s3transfer 0.16.0 pypi_0 pypi
safetensors 0.7.0 pypi_0 pypi
scipy 1.14.1 pypi_0 pypi
setuptools 80.9.0 py310h06a4308_0
sqlite 3.51.0 h2a70700_0
sympy 1.14.0 pypi_0 pypi
tartanair 1.2.0 pypi_0 pypi
tk 8.6.15 h54e0aa7_0
tokenizers 0.14.1 pypi_0 pypi
torch 1.11.0+cu115 pypi_0 pypi
torchaudio 0.11.0+cu115 pypi_0 pypi
torchvision 0.12.0+cu115 pypi_0 pypi
tqdm 4.67.1 pypi_0 pypi
transformers 4.35.0 pypi_0 pypi
triton 3.5.1 pypi_0 pypi
typeguard 4.4.4 pypi_0 pypi
typing-extensions 4.15.0 pypi_0 pypi
tzdata 2025.3 pypi_0 pypi
urllib3 2.6.2 pypi_0 pypi
wheel 0.45.1 py310h06a4308_0
xorg-libx11 1.8.12 h9b100fa_1
xorg-libxau 1.0.12 h9b100fa_0
xorg-libxdmcp 1.1.5 h9b100fa_0
xorg-xorgproto 2024.1 h5eee18b_1
xz 5.6.4 h5eee18b_1
zlib 1.3.1 hb25bd0a_0

result：

[zwdoescode]

@mengxing233-cloud Thanks for trying this out.
Could you please provide your environment? It looks like it got missed in your reply.
BTW, we are planning to provide a build for CUDA 11.x in the near future.

[mengxing233-cloud]

@zwdoescode My computer environment is as follows: Ubuntu 22.04, Python 3.10.19, GPU: Nvidia RTX 4060, CPU: i7-13700K × 24.
I used Conda to configure this project. I installed the CUDA 11.5 environment in the Conda environment, and I confirmed that the project is calling the libraries from the CUDA 11.5 environment when running. I also tested the mask function provided by the project and did not encounter any issues.