The NVIDIA CUDA Deep Neural Network library (cuDNN) is a GPU-accelerated library of primitives for deep neural networks.
The NVIDIA CUDA Deep Neural Network library (cuDNN) is a GPU-accelerated library of primitives for deep neural networks. cuDNN provides highly tuned GPU implementations of standard routines such as forward and backward convolution, pooling, normalization, and activation layers, exposed by a C++ API.
Please use this tag only if you are writing or building code which uses the cuDNN APIs. If you are having problems with installing or running frameworks or libraries which use cuDNN internally, please tag those questions with the library or framework in question, and refrain from tagging them with this tag.
