* add fuse act add grad pass

* polish code

* refine code

* add test

* refine code

* rename some kernel name

* fix compile problem

* fix bug of eigen_dependency

* fix xpu compile

* fix recompute for stop_gradient and inpalce

* fix ut

* update

* delete unnecessary shape and slice op
Co-authored-by: NYour Name <you@example.com>

* clear fluid api: sigmoid_cross_entropy_with_logits

* fix loss.py

* change paddle.nn.functional.sigmoid_cross_entropy_with_logits

* delete sigmoid_cross_entropy_with_logits

* fix binary_cross_entropy_with_logits

* fix ci bug

* fix ci buf

* [Fluid clean]Migrate base/call/print et.al transformer into paddle.jit

* fix phi kernel

* Revert "fix phi kernel"

This reverts commit eff8891c7efda6e49799edbcaef2ca50379d50ef.

some legacy code still use xpu_wait() for stream sync -- it only syncs
default stream. this PR replaces them with dev_ctx.Wait() to ensure
that correct stream is always used

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* [remove fluid] under unittesets

* fix mma_tensorcore (__CUDA_ARCH__)

* disable tensorcore by default.

disable tensorcore by default, because the judgment of __CUDA_ARCH__ will cause undefined behavior in some environments, can manually enable it on a machine that supports tensorcore.

* traspose2 kernel migrated

* Got rid of mutable_data

* x modification added

* ops added in extra info file

* Formatting fix

* 2 fuse passes with tanpose2 commented

* nr of outs changed in 2 passes, passes uncommented

* Changes in passes reverted

* transpose chnaged in operator.cc

* MKLDNN check in operator.cc

* Transpose fixes

* Fix deleted from operato

* template corrected
Co-authored-by: NPaulina Gacek <paulinagacek@intel.com>

* replace LoDTensor with phi::DenseTensor in fluid\operators

* replace LoDTensor with phi::DenseTensor in fluid\operators

* Update split_lod_tensor_op.cc

* Update warpctc_op.cc

* Update broadcast_tensors_op.cc

* Update crf_decoding_op.cc

* Update lstm_op.cc

* Update lstm_op.cc

* Update lod_reset_op.cc

* Update gru_op.cc

* Update linear_chain_crf_op.cc

* resume 2 files for confilct

* Update gru_op.cc

* Update linear_chain_crf_op.cc

* Update lstm_op.cc

* add pattern match

* add unittest

* isort all files

* revert conflicting files

* revert conflicting files

* revert conflicting files

* revert conflicting files

* revert conflicting files

* isort all files

* revert conflicting files

* revert conflicting files

* revert conflicting files

* migrate enforce_custom.h from fluid to phi

* move to backends/custom/

Optimize the implementation of the argsort operator

* [Fluid Clean]Migrate if/while/return/break transformer into paddle.jit

* migrate call_transformer

* migrate call_transformer

* cleanup unused code

* unify is_int8 is_bfloat16

* Simplify matmul_v2 FWD kernel

* remove RunKernel methods

* remove import namespace

* remove headers

* clean fluid/phi cross imports

* remove fluid axpy_handler

* delete fluid methods

* activations

* OneDNNMemDesc

* MKLDNNFormatForSize

* MatchShapeToLayout

* MKLDNNMemoryFormat

* MKLDNNFormat

* ReorderMKLDNNHandler

* to_void_cast

* review suggestions

* interpolate

* remove fluid depedency

* init

* ExecuteMatMulV2

* rm fluid kernel

* matmul_grad

* remove mutable_data

* mul_grad

* matmul fwd

* add extra attr

* temp disable passes

* re-enable passes

* workaround for matmul+act

* fix for matmul+eltwise_add

* fix typo

* merge bugfix #48364

* remove merge conflict

* refine set_state_dict

* fix some typo errors

function rotate: format errors.
function to_grayscale: missing one sentences to explain num_output_chanels

* Update functional.py

* test=docs_preview

* Update functional.py

* Update functional.py

* Update functional.py

* Update functional.py

* Update functional.py
Co-authored-by: NLigoml <39876205+Ligoml@users.noreply.github.com>

* fix:add no support for cuda_arch<700

* replace Executor in while op with InterpreterCore

* cache InterpreterCore as the member of WhileOp

* fix bug: tensor place changed because of assign op in while loop

* refine code

* refine code

* refine code

* hot fix

* fix compile

* merge develop

* follow comments

* add log for test

* remove LoDTensor

* set flag control_flow_use_new_executor false
Co-authored-by: Nfengshuai <fengshuai03@baidu.com>
Co-authored-by: Nzhiqiu <chenqiuliang@baidu.com>

* get default calc stream from execution ctx instead of global dev ctx pool.

* remove lod_tensor_to_array, array_to_lod_tensor, DynamicRNN

* remove less_equal, greater_than, greater_equal, equal, not_equal

* remove pool3d

* Remove API: lod_append
	remove lod_append which is not used in Paddle 2.0

* remove lod_append test file

* add depend

* add origin amp files

* fp16 distinguish None & False

* engine log

* dp add deps for graph exe

* add dep for grad clip

* dep ops in comm stream

* unitest

* move instance_norm from fluid to static

* move group_norm, data_norm to static