* qkv

* ci_test

* add trt LT version helper

* trt8 requires void** to be void* const*

* add trt LT version helper

* remove deprecated nvinfer1::DimsCHW and replace it to nvinfer1::Dims3

* remove deprecated nvinfer1::DimsNCHW and replace it to nvinfer1::Dims4

* update deserialize engine

* update to createNetworkV2

* update to createNetworkV2

* update buildWithConfig and remove redundent config settings

* replace createNetwork to createNetworkV2

* fix int8

* addMatrixMultiply

* remove unnecessary const cast

* IBuilder->setInt8Calibrator() is deprecated

* auto enable fp16 when using int8

* remove the redundant line

[Inference Tensorrt] Add attr for trt engine and handle the input seq problem for ernie var len. (#33575)

* 1, remove layernorm dynamic fp16; 2, let reshape out in dynamic shape

* remove useless code

* add support for trt dynamic shape flatten op

* add version restriction

* add ut input dynamic shape

* Small changes related to BF16 fusion_gru and fusion_lstm

* Correct to pass arg by value

* Add conditions to rnn op

* Correct the spelling mistake

* Improving the test with checking activation

* Trigger CI

* - Second fix

- fix

* - fix

* - First fix to #33021

* add dynamic layer_norm plugin

* fix bug

* fix numpy.allclose

* fix format

* fix code style

* remove shepe in dynamic shape

* code format

* remove layer norm fp16

* fix format

* fix inference prepare data bug

* rename functions

* typo

* typo

* typo

* UT correct

* correct condition

* correct condition

* ci coverage

* morelines

* fix ci coverage

* add trt convert op: reshape

* reshape

* fix

* input: Shape,ShapeTensor

* Add some comments

* download_verify

* checksum

* add more conditions

* dynamic shape

* ut

* correct contidions

* commnent

* remove rebandadnt op type

* remove rebandant if

* revert elementwise and disable trt conv if strides > 1

* strides check

* remove useless var

* comments

* CI skip inference test if only python files modified, test=develop

* fix compilation error on ROCM, test=develop

* fix cmake error on PR-CI-ROCM-Compile, test=develop

* [Paddle-TRT]fix fc_op

* [Paddle-TRT]fix fc_op

* [Paddle-TRT]fix fc_op

* test_trt_subgraph_pass.py

* fix elementwise_op

* fix elementwise_op

* fix elementwise_op

* fix elementwise_op.cc

* op_teller.cc

* implement MHA order same as training

* fix fp16 compile issue on old architecture

* fix format

* fix format

* Add dlnne engine runtime

* Fix log

* Remove <const_cast> and remove unrelated modify with dlnne, +clang-format

* Fix CMakeList format error

* Add copyright message

* Fix dlnne CMakeList.txt

* Add some paddlepaddle_pass to support more networks

* Fix some format bug

* Add delete dropout_op pass

* Fix some format bug

* Fix format bug

* - Added clearing oneDNN per executor

* - Executor is nt always having FLAGS_use_mkldnn set to true

* fix airank bert emb order

* move input num check to converter

* add input num check

* add unused var check white list