Improved test for BilinearSampler

[haojin2]

Description

Fix for #12248.

Checklist

Essentials

• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage:
• Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
• Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
• Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
• Code is well-documented:
• For user-facing API changes, API doc string has been updated.
• For new C++ functions in header files, their functionalities and arguments are documented.
• For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
• Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
• To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change

Changes

• Add cudnn_off option to BilinearSampler
• Remove old flaky unit test for BilinearSampler
• Add new unit test of consistency for BilinearSampler

Comments

CPU, GPU and CuDNN versions all failed on the same seed with the same "wrong" number, which means our 2 versions are consistent with the CuDNN version, so the reference answer calculated by the original helper function within the unit test should be problem here. Changing the test from comparing to the reference answer to a consistency check where the 3 versions are compared against each other to check for correctness.

[haojin2]

New test passed more than 10000 trials on CPU, GPU and CuDNN on both req types:

MXNET_TEST_COUNT=10000 nosetests -s --verbose tests/python/gpu/test_operator_gpu.py:test_bilinear_sampler_versions
[INFO] Setting module np/mx/python random seeds, use MXNET_MODULE_SEED=288629553 to reproduce.
test_operator_gpu.test_bilinear_sampler_versions ... ok

----------------------------------------------------------------------
Ran 1 test in 405.660s

OK

[samskalicky]

LGTM

[haojin2]

@marcoabreu @larroy @lebeg Can you guys take a look when you have time? Should be good for merge.

[haojin2]

@anirudh2290 @sandeep-krishnamurthy @nswamy Should be ready for merge.