* fix safe bug of scatter/scatter_nd

* dygraph sharding

* update unitest hybrid_parallel_communicate_group

* added matmul_v2 bf16/fp32 FWD kernel

added matmul_v2 bf16/fp32 FWD kernel

* added formatting

* removed some tests due to timeout in CI

* refactored tests

* merged tests classes into one file

* minor change

* removed test guard for CUDA

* remove skipIf

* changes after review

* formated one file

* minor change

* added skipping UT in CUDA place

* support set_device for NPU.

* minor update doc and add more unit test.

* scale_matmul_fuse_pass_init

* enhance scale_matmul_fuse_pass

* change scale_matmul_fuse_pass unittest

* add retry in pull sparse

* retry

test=develop.

* remove heterbox

* remove heterbox

* add transpose transpose opdef, test=develop

* add line, test=develop

* fix wrong name, test=develop

* add pass, test=develop

* fix bug, test=develop

* fix bug, test=develop

* delete limite about alpha, test=develop

* add mul to reshape2MatmulFusePass, test=develop

* add limit about alpha, test=develop

* fix bug,test=develop

* set adaptive as false and global_pooling as True, test=develop

* set x_num_col_dims as 1, test=develop

* fix reshape, add attr limit, test=develop

* fix conflict,test=develop

* fix comment, test=develop

* fix comment,test=develop

* fix comment,test=develop

* ,test=develop

* add IsType, test=develop

* add IsType, test=develop

* PyNumber_Long only for numpy, test=develop

* [ROCM] fix RNN miopen as weight need to permuted, test=develop

* [ROCM] fix data share when is_test, test=develop

* update, test=develop

* fix trt flatten converter when batch_size > 1

* change ut to same dynamic shape

* qkv

* ci_test

* in NPU environment, use CPUPlace for missing operators.

* in NPU environment, use CPUPlace for missing operators.

* fix TensorCopy bug and add unit test.

* fix code style.

* add more unit tests.