* add constant_folding_pass pass for mkldnn int8

* update UpdateScaleOpInOutScales

Unified api args name

* add unbalanced data

* fix utest

* fix memory copy in prepare_data. test=develop

* add cache_kv fp16 support. test=develop

* fit for simplify_with_basic_ops_pass. test=develop

Fix english documents of sparse api

* support generating code of opmaker for backward op invoke forward op (#46912)

* [code-gen] Support code-gen for opmaker of sparse op (#46993)

* support generating code of opmaker for backward op invoke forward op

* gsupport code-gen of opmaker for sparse op

* refind logic of choose phi kernrel

* fix complie budg

* fix code_gen bug

* fix bug

* fix kernel signature code-gen

* fix complie bug of VarType

* fix complie bug of VarType

* fix test_sparse_conv_op

* fix test_sparse_norm_op

* [Phi] Refactor logic of judging whether having a phi kernrel (#46920)

* refind logic of choose phi kernrel

* fix complie budg

* update cmake

* update codestyle

* [AutoParallel] fix fp16 for subblock (#47189)

* [AutoParallel] fix fp16 for subblock

* fix engine

* fix comment

* [AutoParallel] fix engine _build and cost method (#47263)

* fix engine build method

* fix import

* update engine cost

* update raise error

* update cmakelist

* revert optimizer

* revert optimizer

* fix unittest

* fix unittest
Co-authored-by: Ncaozhou <caozhou@radi.ac.cn>
Co-authored-by: Ncaozhou <caozhou@radi.ac.cn>

[Cherry-pick][audio] rm kaiser window in audio get_window function && rm audio utils(#47469) (#47479)

* [audio] rm kaiser window in audio get_window function && rm audio utils (#47469)

* rm kaiser window in audio window function

* rm paddle audio utils which is redundant

* rm kaiser in test_audio_functions.py
Conflicts:
	python/paddle/audio/utils/error.py
	python/paddle/tests/test_audio_functions.py

* format

* update dygraph PTQ export_model api

* remove postprocess

* reformat hybrid_parallel_util.py by black

* add fused_allreduce_gradients_with_group

* add scale

* fix ci

* [JITLayer]Enable OneDNN on CPU and Fix zero shape

* [Dy2St]Fix abnormal growth of memory in train mode and no_grad for Dy2St 

* [JIT] Add Predictor for JITLayer (#47379)

* add predictor_engine

* add predictor_engine

* fix zero shape

* fix lodTensor

* fix unittest

* fix code style

* update CmakeList

* fix new executor

add sync_batch_norm_bn and deliver indices_dict 

[Cherry-pick Release/2.4] Fix multi_tensor adam and momentum bug when the parameter is list of dict (#47372)

* reformat file by black

* fix multi_tensor adam/momentum bug

* cherry-pick #46359 and resolve conflict

修改一处Slice的bug

* fix inference perfermence problem caused by selecting cudnn kernel for softmax

* recover use_cudnn in opmaker of softmax

* fix tess split fold

* format

[Dy2Stat]Fix module loading OSError in multiprocess

* return proper state

* fix for dims

* fix
Co-authored-by: Njakpiase <jakpia21@gmail.com>

[Cherry-pick][Release/2.4]Refine the memory usage of fused_attention and fused_feedforward ops (#47235)

* fix fused_attention fused_feedforward

* fix ci

* fix ci

* fix ci PADDLE_GET_CONST

* fix ci ut

当前无法从Tensor中获取到SparseTensor的sparse_dim，无法准确推断出indices的shape，所以目前先以3D点云模型为主，输入的SparseTensor的维度是5D的，其中非零元素是一维向量，所以indices是[4, -1]。

* fix geometric doc

* add prior_box and box_coder for paddle.vision.ops

* fix UT change assertTrue to assert_allclose

* fix formula format

* [CodeStyle] add black config to release2.4

* empty commit, test=document_fix

* Fix hAPI bug of not compatible with LayerHook

* Fix bug of reduce_sum op. When input.numel() > INT32_MAX, its result
is wrong.

* support pure bfloat16

* support bf16 linear

* update PR to pass CI

* tiny fix where_grad_kernel.cu

* Support bfloat16 type for reducer and sharding.

* Fix some bug.

* Polish code.

* Polise code.

* Add bfloat16 datatype in fill_grad kernels.
Co-authored-by: Nsneaxiy <sneaxiy@126.com>
Co-authored-by: Nsneaxiy <sneaxiy@126.com>

* forbid Conv2D into Paddle-TensoRT when weight is not persistable.