* Add unittest redist holder for all test projects.

* Move redist holder to another folder.

* Create dotnet.yml to config github action

* Revise the sln file.

* Update dotnet.yml

* Add version to tensorflow.hub.

* Create release.yml

* Update dotnet.yml

* Update and rename dotnet.yml to build_and_test.yml

* Update release.yml

* Update release.yml

* Revise project using.

* Update build_and_test.yml

* Update release.yml

* Update the package info of Tensorflow.Hub.

* Add a tolorance to equivalence of NDArray.

* Create semantic.yml

* fix: run code clean.

* Update release.yml

* Update release.yml

* ci: revise the auto release ci.

* ci: update release ci.

* ci: update release ci.

* ci: update ci files.

* ci: update ci files.

* ci: update ci files.

* ci: update ci files.

* ci: update ci files.

* ci: update release ci and hub package info.

* ci: revise build_and_test ci.

* test: add tolorance to float NDArray comparison.

* ci: disable linux test.

* ci: update release ci.

* Update release.yml

* ci: update release ci.

* ci: revise auto-release ci.

* ci: update auto-release ci.

* Update release.yml

* ci: specify packed project names of auto-release.

* ci: revise auto release ci file.

* ci: revise auto-release ci file.

* ci: revise auto-release ci file.

* ci: revise auto-release ci file.

* Update release.yml

* ci: revise auto-release ci file.

* ci: revise auto-release ci file.

* ci: revise auto-release ci file.

Add the constructor of NDArray which reuses memory

Partially fix the error when crop image.

Fix (#1036) and adjust the keras unittest

Change some project information

Add Tensorflow.NET.Hub

Change type of BuildInputShape and BatchInputShape

Fix the error of loading model saved before tf2.5.

Add set_weights and get_weighst APIs

Fix the error when loading Conv1D layer with initialzier.

Fix the error of example caused by #1022

Fix the error when loading VGG19.

Partially Support the function loading

Fix the error when saving model with GPU.