Blog 2: CV-CUDA is Now on PyPI – Install with pip | CV-CUDA at GTC 2025

[shiremathNV]

Hello CV-CUDA community!

We are excited to announce that CV-CUDA v0.14 is now available on PyPI.org, making it easier for developers to discover, install, and integrate GPU-accelerated computer vision into their Python workflows. With this release, Python wheels for x86 architectures are now accessible via pip, simplifying the installation process.

Key Updates:
CV-CUDA v0.14 x86 Python wheels are now on PyPI.
Developers can install the library using:

pip install cvcuda-cu12 # For CUDA 12.x users
pip install cvcuda-cu11 # For CUDA 11.x users

Available packages:
cvcuda-cu12 – For CUDA 12.x users
cvcuda-cu11 – For CUDA 11.x users

NOTE:

• Starting with v0.14, aarch64_cu12 packages (deb, tar.xz or wheels) distributed on GitHub (release "assets") and Pypi are the SBSA-compatible ones. Jetson builds (deb, tar.xz, whl) can be found in explicitly named "Jetson" archives in GitHub release assets.
• We do not provide SBSA-compatible aarch64_cu11 packages yet, this will be addressed in an upcoming release.

In addition to PyPI availability, there were other new features added in the v0.14 release. Please refer to the release announcement for details.

Why Use CV-CUDA?

• Open-source library for Python developers (includes C++ APIs)
• High-performance GPU acceleration for large-scale image and video processing.
• Optimized for AI pipelines, handling millions of images per second with CUDA-accelerated operators.
• Seamless integration with PyTorch, TensorFlow, and AI inference workflows.

Get Started
Install the library today and improve your computer vision processing with GPU acceleration. For further details, refer to our documentation and developer resources.

Resources
CV-CUDA Documentation
PyPI: cvcuda-cu12 | cvcuda-cu11
Customer Story: Optimizing Microsoft Bing Visual Search with NVIDIA Accelerated Libraries

Attending GTC 2025?
If you are attending GTC next week, do not miss these sessions that will be highlighting CV-CUDA among other acceleration libraries used to accelerate image and video processing workloads for developers-

• Accelerating AI Pipelines: Boosting Visual Search Efficiency [S71676]
• Optimize Short-Form Video Processing Toward the Speed of Light [S73178]
• Meet with Experts: Visual Data Preprocessing Using NVIDIA Libraries/SDKs [CWE72379]