* Support parallel test for understand_sentiment

* Full test on understand_sentiment

* Skip normal tests

* Debug CI

* Enable benchmark

* Revert init.cc

* Make CI pass

* Generating random numbers with given batch size

uniform_random_batch_size_like_op
gaussian_random_batch_size_like_op

* More comments about random seed.

* Move test_*_random_batch_size_like_op to unittests

Currently, our tests run with 2 GPUs, the init time is absurdly long:
about 4s for each process.  Currently, we run each OP test on
different processes. This PR:

1. create cmake function py_test_modules which will generate the
Makefile that runs a list of Python unittest module in a single Python
process.

2. move all "python unittest compatible" (e.g., used the unittest
package, not just a regular python file). from fluid/tests to
fluid/tests/unittests.

3. cmake now will run all OP tests in fluid/tests/unittests in a
single process, except the time-consuming tests, they are separated
into different processes to utilize parallelism. Please make sure to
use the unittest package if you put the python test file in
fluid/tests/unittests

4. remove all exit(0) from fluid/tests/unittests/*.py, exit(0) is used
to disable unittest, we can not do it when running all tests in a
single process since it will terminate the process without running the
other tests. Instead, the test is disabled in
fluid/tests/unittests/CMakeLists.txt. FIXME is added for each disabled
item. Please disable the unittest from
fluid/tests/unittests/CMakeLists.txt, instead of adding exit(0) to the
Python file, for all Python file in fluid/tests/unittests/.

5. add an option WITH_FAST_BUNDLE_TEST. When OFF, will run the unit
tests in separate process so that they can be tested individually.

* Implement multibox loss wrapper in Python API.

* Add some wrappers for SSD detection.

* Fix conflicts.

* Add unit testing for SSD loss wrapper.

* Update doc in Python API.

* Refine unit testing.

* Add more unit testing and update some interface arguments.

* Add softmax_with_cross_entropy and smooth_l1 in Python API.

* Fix doc format.

* dist train support lr decay

* update by comment

* revert elementwise method creator

* delete comment

* Refine target_assign_op to unify the classification and regression targets assignment.

* Fix the unit testing.

* Fix conflicts.

move paddle/pybind/pybind.h to paddle/fluid/pybind/pybind.h, and cancel the test_parallel_op temporary

* Added an elementary test case for CSP

* removed input

* Rename test file to avoid running in CI

* Fix YAPF error

* Remove one line function handler

* Add unit-test and example

* Fix type error

* Fix unit test cases

* Fix init error for cudaplace

* Change unit-test options

* initial commit

* fix bug

* end of file fix

* address comments

* initial commit

* minor fix

* remove redundency

* address comments

* Save the inference model in Python example of recommender_system.

* Add infer() in Python unittest recommender_system.

* Add C++ inference unittest of recommender_system.

* Refine load

* Address review comments: round 1

* Make API consistent with python-save/load

* Add another unit test

* Remove commented function

* Fix GPU bug

* Address review comments

* Modify wrt PR 8147

* Fix filenames for combined case

* Fix typo

* Address review comments: round 2

* Unify TestInference by keeping default param in template

* Address review comment

* Fix spacing