It also includes higher level SDKs such as NGX SDK which makes it easy to integrate AI features into creative applications with pre-trained networks. For example libraries like cuDNN and DALI help developers reduce training times while TensorRT optimizes networks and takes advantage of Tensor Cores for the fastest inference on GPUs. These libraries accelerate the entire AI pipeline from data processing to training and inference to deployment. CUDA parallel programming model for general computing on GPUs.ĬUDA-X AI is a collection of optimized libraries built on CUDA.Capture SDK capture and compress the desktop buffer for transmission or storage.

GPUDirect for Video efficiently transfer video frames in and out of NVIDIA GPU memory.

Optical Flow SDK highly accurate flow vectors, with robust frame-to-frame intensity variations and true object motion for tracking objects within video frames, video action recognition, stereo depth estimation and many more applications.

Video Codec SDK hardware accelerated video encode and decode of industry standard formats.

Nvidia studio drivers