This new release provides significant efficiency gains by leveraging the optimized data processing of PatchTensor and inference of the PatchGNP. This uses the new separable, block-factorized kernels (see papers). This greatly improves the average running times (5x+ or more) from version 1.0.0 to 2.0.0 on both CPU and CUDA devices. New examples for training are also included in this release.