Support backward of backward for Relu and add a new gradient checker by comparing theoretical and numerical Jacobian. (#16862)

* Support backward of backward and a new gradient checker
* Rename decorators.py to decorator_helper.py, since Python on Windows CI has decorators package.

1. Add ReluDoubleGradMaker when register relu_grad.
2. Add a new gradient checker by comparing theoretical and numerical Jacobian.  Check double gradients by double_grad_check.

* move gc test to op_test
test=develop

* Revert "move gc test to op_test"

This reverts commit cf15da65.

* enable gc test in some ops
test=develop

* fix ramdom fail
test=develop

test=develop

* add parallel build script to ci test=develop
* 1. classify the test case as single card/two cards/multiple cards type
   2. run test case according to the run type

test=develop

* Speed affine_channel_op unit testing
* Add check in tensor_py
* Fix ONLY_CPU Compiling

Update the filter generation mechanism that it could generate the negative parameter.
The original calling(np.random.random()) couldn't simulate the conv/relu fusion case.

test=develop

* accelerate test_ir_memory_optimize_nlp, test=develop

* accelerate test_ir_memory_optimize_nlp, test=develop

Rename the testcuda function to has_cuda, it will elimate the unnecessary testing.
test=develop

* speedup gc and inplace softmax_with_cross_entropy_grad
test=develop

* refine models gpu mem
Merge skip vars and warning messages of mem opt
remove relu mem opt
test=develop

* follow comments
test=develop

test=develop

* Split test_calibration test to two tests

test=develop

* Modify CMakeLists file
test=develop

* Support seq len equal to 0 in sequence ops

test=develop

* Add more test cases

* Fix some comments

test=develop

* Fix py3 error

test=develop

* Check some shapes only in runtime

test=develop

* Follow review comments

test=develop

* Update API spec

test=develop

test=develop

* fix sampling id op bug, test=develop

test=develop

test=develop

test = develop
resolve #16764

* test=develop
update API.spec with hash, resolve conflict

* remove unused parameter

* optimize lstmp and sample_logits op, test=develop

* update op_use_default_grad_op_maker.spec, test=develop

* delete useless file,test=develop

* append 0 to dim variable to avoid memory reusage, test=develop

* test=develop, fix hsigmoid dereference nullptr

* test=develop, refine condition

* test=develop, refine comments

test=develop

test=develop

test=develop

test=develop