Release 1.1.0

Changes

1. Fix AVX2/AVX512 and non-AVX x86 kernel for Op=T
2. Better performance (upto 20%) on CUDA when OpX=T or OpF=T for MKM

v1.0.1

Changes

1. Do gradient computation only for tensors that requires_grad.

Welcome to FastKron

The first release of FastKron adds support for x86 CPUs and NVIDIA GPUs. PyFastKron supports both PyTorch tensors and NumPy arrays. This release contains PyFastKron for CPython 3.9 to 3.12.