Weaver of Adaptive Vision – a Zebra Technologies company – is a high performance inference engine for machine vision. It executes deep neural networks on both Nvidia GPUs and Intel CPUs with the highest performance. The library is coming with both C and modern C++ interfaces. Users need to convert their data structure to weaver’s tensor (multi-dimensional array) and then can invoke its two functions: deploy and run. Most of the other inference engines require to do Python programming and tweak many things. Weaver provides 3-10 times better execution time than an open-source framework without optimizations.
