Accelerated Learning cuDNN provides kernels, targeting Tensor Cores, to deliver best available performance on compute-bound operations. It offers heuristics for choosing the right kernel for a given problem size. Expressive Op Graph API The user defines computations as a graph of operations on tensors. The cuDNN library has both a direct C API and an open-source C++ frontend for convenience. Most
![NVIDIA cuDNN | NVIDIA Developer](https://cdn-ak-scissors.b.st-hatena.com/image/square/c508c975172534ac2931db7e4309e753fccf91bb/height=288;version=1;width=512/https%3A%2F%2Fdeveloper.download.nvidia.com%2Fimages%2Fog-default.jpg)