Initial commit for dataset op python

Added signature to barrier

Adde compiling barrier code

Rebasing, fixed new compile errors

Final fix for make_unique

Added pybind API for barrier

Fixed pyfunc invocation

python interface - sync_wait

!1 sync_wait python interface
* python interface - sync_wait

fix test

update test

update test

Added new test case

add test case

test for shuffle + batch

Added two-sync test case

Restrited that no shuffle after sync

Added sync to pipeline info

block first databuffer as well

Intelligently get batch size

Fix default case

Lock Pair shares among all iterators

Added fix for empty character

Fixed up test case formatting

Fix end of epoch in sync_wait

Fixing CI

addressed code review comments. added check in python layer to exclude directories and to raise an error if a pattern does not match any file

fixed clang format

fixed cppcheck

fixed cppcheck (used std::accumulate and std::copy_if). regenerated tfrecord file to contain correct header, it was a dummy header before

fixed cppcheck: added const reference for string parameter for lambdas, fixed clang format: whitespace adjustments

more clang whitespace fixes...

changed print to logger.info

ALLGATHER, and REDUCESCATTER with different factors; change the BETA and
GAMMA value in cost model.

1) A[B]=U
2) A[A>n]=U
    A.shape == B.shape
    U is a scalar or Tensor(size==1)
    B is Tensor(dtype=bool)
    n is a Number
Signed-off-by: Ncandanzg <zhangshucheng@huawei.com>

Signed-off-by: Ncandanzg <zhangshucheng@huawei.com>

clean clang format errors and cpplint errors

add some test cases for histogram summary operator

Add new test cases

format code

remove test case for fp64

add back scalar manipulation

remove save graph

fix pylint