NVIDIA が CUDA-X AI ライブラリを更新
CUDA-X AI ソフトウェア高速化ライブラリは、現代の AI アプリケーションにおける GPU の能力を解き放ちます。 新しいアップデートにより、Tensor コアの活用、NLP 向け BERT の高速化、音声駆動のレコメンド エンジンの構築等が可能になります。
NVIDIA Announces CUDA-X AI SDK
March 18, 2019
At GTC Silicon Valley in San Jose, NVIDIA released CUDA-X AI, a collection of NVIDIA’s GPU acceleration libraries built on CUDA that accelerate deep learning, machine learning, and data analysis.
CUDA-X AI includes cuDNN for accelerating deep learning primitives, cuML from RAPIDS.ai for accelerating machine learning algorithms, NVIDIA TensorRT for optimizing trained models for inference, and over 15 other libraries. Together, they work seamlessly with NVIDIA Tensor Core GPUs to accelerate the end-to-end workflows for developing and deploying AI-based applications.
CUDA-X AI is integrated into all deep learning frameworks, including TensorFlow, Pytorch, and MXNet, and leading cloud platforms, including AWS, Microsoft Azure, and Google Cloud.
CUDA-X AI libraries are freely available as individual downloads or as containerized software stacks for many applications from NGC. They can be deployed everywhere on NVIDIA GPUs, including desktops, workstations, servers, cloud computing, and internet of things (IoT) devices.
CUDA-X AI ソフトウェア高速化ライブラリは、現代の AI アプリケーションにおける GPU の能力を解き放ちます。 新しいアップデートにより、Tensor コアの活用、NLP 向け BERT の高速化、音声駆動のレコメンド エンジンの構築等が可能になります。
NVIDIA Announces CUDA-X AI SDK
March 18, 2019
At GTC Silicon Valley in San Jose, NVIDIA released CUDA-X AI, a collection of NVIDIA’s GPU acceleration libraries built on CUDA that accelerate deep learning, machine learning, and data analysis.
CUDA-X AI includes cuDNN for accelerating deep learning primitives, cuML from RAPIDS.ai for accelerating machine learning algorithms, NVIDIA TensorRT for optimizing trained models for inference, and over 15 other libraries. Together, they work seamlessly with NVIDIA Tensor Core GPUs to accelerate the end-to-end workflows for developing and deploying AI-based applications.
CUDA-X AI is integrated into all deep learning frameworks, including TensorFlow, Pytorch, and MXNet, and leading cloud platforms, including AWS, Microsoft Azure, and Google Cloud.
CUDA-X AI libraries are freely available as individual downloads or as containerized software stacks for many applications from NGC. They can be deployed everywhere on NVIDIA GPUs, including desktops, workstations, servers, cloud computing, and internet of things (IoT) devices.