* add fused_multi_transformer_encoder/decoder pass, run GPT-3 success

* first version, accuracy corrected

* disable debug print

* use blockReduceSum in phi

* add UT

* add opCompat

* code style

* code refine

* bug fix

* code refine

* test fix

* bugfix

* codesytle fix

* code style

* code-style

* code-style

* code-style

This reverts commit 166ff39a.

* first version of ln_s_p with s>0

* refine and UT

* pass opt draft

* pass opt

* code refine

* code-style

* bug fix

* fix ci test

* code style

* [Paddle Inference]support n lookup_tables fuse to embeddinglayernorm(3)

* Remove trt_reshape2_matmul_fuse_pass

* first commit

* conver done

* correct format

* layernorm_shift_partition

* correct convert

* redefine plugin

* runable

* bug fix

* modify ShiftPartitionPattern

* correct

* add UT

* modify ut

* compile

* modify enforce

* modify UT

add constant folding pass， for some model，it will get less latency；

* Add int8 support for matmul+elementwiae_add fuse

* Corrections after review and ernie test fix

* merge conv_concat_relu to conv_act

* fix typo

* extend unit test

* reuse existing gpd

* codestyle

* enforce mkldnn conv

* convert multihead to oss

* fix:bug

* fix:delete const cast

* fix:don't support bias_qk

* add vit pass

* fix:convert bug and add preln_residual_bias

* support length=-1

* add UT for convert

* add no_bias_qk support for gpu_multihead_op

* delete infer_shape depends on bias_qk

* oss just can be used in T4 and A*

* fix:change api for ROCM CI

This reverts commit 84bf5c31.

* cuda_graph

* cuda_graph_

* cuda_graph_

* cuda_graph_

* remove v2_transpose_reshape

* matmul_transpose_reshape

* reshape_transpose_matmul

* restore ut

* adjust old ut

* restore parallel UT ruels

* feedback from review

* Add unit tests

* matmul_v2 + activation

* matmuls + elementwise_add

* matmul_v2 postops

* transform matmul to v2

* opcompat

* fix fusing matmul with multipe outs

* add shape constraints

* remove unused vars

* change pass order

* - Unit tests to be debugged

- fix

- refactor

- diagnostic

- more diagnostic

- fix

- Fix number two

- fix

- fix

- fix

- alpha added

- more fixes

- compilation fix

- removed diagnostic code

- cosmetic fixes

* lint

* add alpha constraint

* merge matmul refactor

* trigger CI

* - fix

* - another fix

* code style

* add support for matmul+elementwise_add+activation

* code style

* fix bfloat16 bugs

* change append_binary to append_sum
Co-authored-by: NJacek Czaja <jacek.czaja@intel.com>

* multihead matmul add fp16

* fix windows error

* fix rocm error

* fix rocm error

* fused_fc_elementwise_layernorm support fp16

* fused_fc_elementwise_layernorm support double

* fc support fp16

* add a ‘,’ on paddle_pass_builder.cc

* fc support fp16 on non-cuda.

* convfusionfp16

* convfusionfp16

* convfusionfp16

* add method for post ops

* format code

* gpd

* format style

* add matmul+act test

* implement matmul+activation

* whitespaces

* code style

* python code format

* Increase UT timeout

* code format

* update style

* generalize activation fuse passes

* change order

* Unify activation GPD

* Revert changes with op_act

* remove softmax mkldnn attrs

* set common name for act attributes

* whitespace

* append postops by helper function

* ut style

* revert changes related to quantization

* Reduce redundancy

* reduce number of parameters

* trigger CI

* validate attribute

* trim unit test

* inference add convert to mixed model ability.

* revert 40531

* update

* fixed multihead matmul fuse pass

* Add unittests

* rm scale op

* fixed code style

* fixed code style

* resolve testcase falied

* add note

* consolidate conv act passes

* generalize conv_activation

* integrate conv+act tests

* code style format

* whitespaces

* remove timeout from old tests

* implement comments from review

* restore ut

* whitespace

* code style

* transpose

* fixes after review

* method for gettin act

* Change Paddle_enforce error type

* code format

* add missing opcompats

* Enable fc on bfloat16

* Add pass for residual connection

* Dissable Residual connection pass for now

* Ban ResidualData from DQ

* style

* WO for python tests

* new general transformer inference support

* support yolov5s static/int8

* fix eltwise_sub and div weight compute

* fix delete_fill_constant_pass

* [Paddle-Inference] support transformer generation: some passes

* Readd conv_affine_channel fuse pass as mkldnn pass

* Fix formatting

* Add new test to parallel_UT_rule.py

* Fix Coverage and Windows CI issues

* Revert "Fix Coverage and Windows CI issues"

This reverts commit f33459846385c9fd51c07f9f44e7ff283a652637.

* Fix CI errors

* Remove unnecessary conv_eltwise_add_affine_channel fuse pass

* Remove test from parallel_UT_rule.py