CV-CUDA is an open-source project that enables building efficient cloud-scale Artificial Intelligence (AI) imaging and computer vision (CV) applications. It uses GPU acceleration to help developers build highly efficient pre- and post- processing pipelines. CV-CUDA originated as a collaborative effort between NVIDIA and ByteDance.
Refer to our Developer Guide for more information on the operators avaliable as of release v0.2.0-alpha.
To get a local copy up and running follow these steps.
- Linux distro:
- Ubuntu x86_64 >= 18.04
- WSL2 with Ubuntu >= 20.04 (tested with 20.04)
- CUDA Driver >= 11.7 (Not tested on 12.0)
- GCC >= 11.0
- Python >= 3.7
- cmake >= 3.22
The following steps describe how to install CV-CUDA from pre-built install packages. Choose the installation method that meets your environment needs.
tar -xvf nvcv-lib-0.2.0-cuda11-x86_64-linux.tar.xz
sudo dpkg -i nvcv-lib-0.2.0-cuda11-x86_64-linux.deb
pip install nvcv_python-0.2.0-cp38-cp38-linux_x86_64.whl
Follow these instruction to successfully build CV-CUDA from source:
-
Build CV-CUDA
cd ~/cvcuda ci/build.sh
This will compile a x86 release build of CV-CUDA inside
build-rel
directory. The library is in build-rel/lib, docs in build-rel/docs and executables (tests, etc...) in build-rel/bin.The script accepts some parameters to control the creation of the build tree:
ci/build.sh [release|debug] [output build tree path]
By default it builds for release.
If output build tree path isn't specified, it'll be
build-rel
for release builds, and build-deb for debug. -
Build Documentation
ci/build_docs.sh [build folder]
Example: `ci/build_docs.sh build
-
Build Samples
./ci/build_samples.sh [build folder]
(For instructions on how to compile samples outside of the CV-CUDA project, see the Samples documentation)
-
Run Tests
The tests are in
<buildtree>/bin
. You can run the script below to run all tests at once. Here's an example when build tree is created inbuild-rel
build-rel/bin/run_tests.sh
-
Run Samples
The samples are installed in
<buildtree>/bin
. You can run the script below to download and serialize the model and run the sample with the test data provided../ci/run_samples.sh
-
Package installers
From a succesfully built project, installers can be generated using cpack:
cd build-rel cpack .
This will generate in the build directory both Debian installers and tarballs (*.tar.xz), needed for integration in other distros.
For a fine-grained choice of what installers to generate, the full syntax is:
cmake . -G [DEB|TXZ]
- DEB for Debian packages
- TXZ for *.tar.xz tarballs.
CV-CUDA is an open source project. As part of the Open Source Community, we are committed to the cycle of learning, improving, and updating that makes this community thrive. However, as of release v0.2.0-alpha, CV-CUDA is not yet ready for external contributions.
To understand the process for contributing the CV-CUDA, see our Contributing page. To understand our committment to the Open Source Community, and providing an environment that both supports and respects the efforts of all contributors, please read our Code of Conduct.
CV-CUDA operates under the Apache-2.0 license.
CV-CUDA, as a NVIDIA program, is committed to secure development practices. Please read our Security page to learn more.
CV-CUDA is developed jointly by NVIDIA and ByteDance.