…dle#35881)

* cherry-pick 32842

…s/Mac (PaddlePaddle#35897)

ATT, cherry-pick PaddlePaddle#35690

…ERTIES TIMEOUT (PaddlePaddle#35863) (PaddlePaddle#35898)

Increase test_imperative_auto_mixed_precision PROPERTIES TIMEOUT from 120s to 300s.

…r python3.6 (PaddlePaddle#35862) (PaddlePaddle#35900)

fix bug of module paddle has no attribute fluid for python3.6.

…le#35882)

….6 (PaddlePaddle#35848) (PaddlePaddle#35874)

* fix bug

…void many compiling warnings. (PaddlePaddle#35839) (PaddlePaddle#35868)

Cherry-pick PaddlePaddle#35839

PaddlePaddle#35879) (PaddlePaddle#35907)

)

Co-authored-by: joanna.wozna.intel <[email protected]>

…#35926)

* Pass compat of conv_transpose_bias_mkldnn_fuse_pass

* Fix a bug of strided_slice op, about the axes parameter access memory out of bounds

* Fix a bug of transpose op, about accessing memory out of bounds of the perm param

* op:transpose_op supports bool type

…addle#35812) (PaddlePaddle#35919)

cherry-pick PaddlePaddle#35812，修复Eigh OP

…ase/2.2 (PaddlePaddle#36015)

解决Windows中CUDA11.2编译出错的问题。
cherry-pick PaddlePaddle#35941

* update xpu version

Add basic Cost Model, it uses executor to run program and profile it to get op time.

This is an early basic version, we will add more functions in the future.

…PaddlePaddle#35909) (PaddlePaddle#36038)

This PR implements the kernel of "eigvals" OP with the Lapack library, which has a better performance than the previous Eigen library.

)

…dle#35934)

(cherry picked from commit 347b182)

…ddle#35927)

1、Split function GradScaler::minimize() to GradScaler::step() + GradScaler::update()
2、Add GradScaler::unscale_(optimizer)

* fix pad tuple

* fix format

* add randperm_op_npu

* fix test_set_value_op_npu

…Paddle#36056)

This PR supports linalg.solve calculation for linear algorithm module of Paddle. One may call paddle.linalg.solve to use it.

…e API PaddlePaddle#36024

[cherry-pick] Add function comments and instructions to the Primitive API

)

This PR added det and slogdet API to release/2.2
It is cherry-pick from PaddlePaddle#34992 and PaddlePaddle#36013

…dlePaddle#35916) (PaddlePaddle#36091)

cherry-pick PaddlePaddle#35916，CPU前向计算将Eigen替换为Lapack，修改linalg暴露规则

…6096)

…) (PaddlePaddle#36094)

When users use gumbel_softmax, they can use paddle.seed() in python for fixed seed.

…36028) (PaddlePaddle#36103)

The AdamW optimizer modify the op from adamw to adam in pr35521, this is a inappropriate modify. Modify adam to adamw in AdamW.

cherry-pick PaddlePaddle#35858、PaddlePaddle#35895

…36111)

cherry-pick from PaddlePaddle#35352

Add new detection api paddle.vision.ops.psroi_pool and paddle.vision.ops.PSRoIPool

…Paddle#36133)

ATT, cherry-pick PaddlePaddle#36098

remove recent linalg api in paddle.init;
add args 'name' in some new linalg api interface

…Paddle#36163)

* fix unique unstack dim 0

* fix unique_op format

* Initial Commit

* fix py2 error

* fix wrong words and doc

* test=document_fix

* fix _gpuDeviceProperties

…e#36160) PaddlePaddle#36161

ATT, cherry-pick PaddlePaddle#36160

…addlePaddle#36131)

根据线性代数库的API暴露规则修改multi_dot的API暴露规则：
1、在python/paddle/tensor/linalg.py 路径下实现
2、在python/paddle/linalg.py 下import并加入__all__列表
3、在python/paddle/tensor/init.py下引入并加入tensor_method_func列表
4、删除了pythonpaddle/init.py的import

* add roi pool

* rename input as x

…ePaddle#36174)

* test=document_fix

* test=document_fix

* test=document_fix

* test=document_fix

…device_capability. (PaddlePaddle#36172)

* add get_device_name and get_device_capability

* fix docs

* fix docs

* fix decs

向PaddlePaddle中的线性代数库添加eig算子，该算子计算一般方阵的特征分解。
cherry-pick 自PaddlePaddle#35674.

add roi align, cherry-pick PaddlePaddle#35102

…PaddlePaddle#36123) (PaddlePaddle#36193)

…#36211) (PaddlePaddle#36233)

* fix raw optim

* pre-commit test file

Co-authored-by: sneaxiy <[email protected]>

Co-authored-by: sneaxiy <[email protected]>

* update func name

* skip cpu

* update unittest

* update unittest

…or (PaddlePaddle#36294)

对于__getattr__重载后不满足条件的参数，全部抛出AttributeError异常，达到与未重载版本一致。

(cherry picked from PR PaddlePaddle#36229)

(cherry picked from PR PaddlePaddle#36095)

PR主要功能：支持C++开发注册GeneratePass，简化针对fusion等子图优化场景开发方式。

…36353)

* Fix stop_gradient in RunProgramOp

* fix reference

* change the paddle.mm to matmul_v2

* update the code for the mm

* update the document for the mm

…PaddlePaddle#36375)

* change time to remove static tempfile

* delete remove_static_file() function

…addlePaddle#36227) (PaddlePaddle#36370)

ATT，cherry-pick PaddlePaddle#36227

* add sparse_embedding doc

* modify sample code

* fix sample code error

…PaddlePaddle#36453)

* [WIP]Verify the correctness of graph rewrited by GeneratePass, test=develop

* add delete subgraph and unittest, test=develop

* check simple pass, test=develop

* fix coverage, test=develop

* limit with input_spec via Paddle API, test=develop

…ing save/load (PaddlePaddle#36434) (PaddlePaddle#36463)

修复使用jit.save/load接口加载模型后，在train模式和no_grad上下文中，显存会一直增长的问题

…lePaddle#36501)

* fix async_read bug

* change index place to cpu

* add tensor size judge

* add async_read & async_write test

* fix bug in async_write

* fix mac py3 ci

* fix bug for cpu version paddle

* fix windows ci bug

* change input argument error type

* change const_cast to mutable_data

* add async_write out-of-bound check and consumate error hint

* fix a small bug for dst_tensor

* add docs and refine codes

* refine docs

* notest,test=windows_ci

* fix windows ci

* fix require

* fix code-block

* add core.is_compiled_with_cuda()

* quant support matmul_v2

* fix format

…ePaddle#36500)

    The code of this PR can only support CUDA 11.2. Currently, CI does not have GPU with CUDA 11.2 , and all tests will be skipped automatically.

    The new OP is paddle._C_ops.sparse_attention. Regarding the work of the python API, it will be resolved in a follow-up PR.

    The code of this PR lacks tests on dynamic graphs and static graphs, and will be added in subsequent PRs.

…addlePaddle#36536)

* catch the generatorfunction and intercept it.

* add test generator

* add test case

* refine the testcase

…#36552)

* remove no_value using var.name

* fix replicate pad when input size is 0

* add unit test

…dlePaddle#36418)

* Add functor_primitives.h for kernel primtive api

…ddlePaddle#36373) (PaddlePaddle#36616)

* Fix a bug in ReadData, ReadDataBc and ReadDataReduce when NX != 1
* Update the implement of reduceAnyKernel according to kernel primitive api

* add viterbi decode cpu kernel

* add viterbi decoder api in paddle.text

* add a data buffer once to avoid create many small pieces of data buffer frequently

* fix viterbi max_seq_length bug

* fix seq_len=1 bug

* fix device context

* move split out of for loop

* remove INVERSE_SUB

* remove 2 GET_CAST_MASK

* remove 1 loop

* remove Functor

* add to_static deploy code

* use MAX_FUNC instead of ELE_MAX

* add MaxFunctor

* impl max_func

* remove MaxFunctor

* remove cast op

* use REGISTER_OP_WITHOUT_GRADIENT

* add viterbi cuda kernel

* add FIX_BLOCKDIM_CASE macro

* add MKL add, mul; add get data mask

* add arange mkl impl

* add CPU Argmax

* add cpu gather

* use EXECUTE_MKL_ELEMENT_BINARY_OP instead of some ADD, MUL

* use SameDimsBinaryOP instead of EXECUTE_MKL_ELEMENT_BINARY_OP

* use SAME_DIMS_ELEMENT_BINARY_OP

* add SimpleBroadcastBinaryOP

* use int instead of int64_t to accelerate

* optimize SimpleBroadcastBinaryOP

* optimize SimpleBroadcastBinaryOP

* optimize performance in both single thread and multithread situation

* remove useless line

* remove useless code

* add CREATE_TENSOR_BUFFER macro

* add INIT_REQUIRED_TENSOR macro

* add comment

* fix windows ci

* add viterbi unittest

* remove cuda add functor

* remove cuda equal

* remove a template function

* fix windows ci

* fix windows dtype

* remove some template instance

* remove useless header file

* remove some blockdim

* remove transpose impl

* accelerate cpu performance on single thread situation

* viterbi_decode->crf_decode

* rename crf params name

* add viterbi api test

* remove useless import

* add enable_static

* use viterbi decoder

* fix viterbi len=1

* fix  viterbi unittest

* remove useless comments

* reconstruct viterbi decode

* remove ADD,SUB,MUL structure

* fix coverage

* remove CREATE_TENSOR

* add name args

* crf.py->ops.py; with_start_stop_tag->include_start_end_tag

* update crf_decode en docs

* fix viterbi decode en docs

* fix some review comments

* add FIXED_BLOCK_DIM_CASE in cuda

* push_back->emplace_back

* crf_decode->viterbi_decode; include_start_end_tag->include_bos_eos_tag

* paddle.text.ops.viterbi_decode->paddle.text.viterbi_decode

* fix viterbi_decode en docs

…ddle#36640)

In fused_attention op and fused_ffn op, the fused bias_add+dropout+residual+layernorm kernel or bias_add+dropout+residual kernel is used. To ease the use of this kernel, we provide a wrapper in this PR.
1.To reuse the increment computing code, we exact the corresponding code to "GetSeedDataAndIncrement" routine in dropout_impl_util.h.
2.The fused_dropout_helper.h provides the fused dropout kernel wrapper.

Note: the test of this warper will be provided in the following fused_attention_op and fused_ffn PRs.

…lePaddle#36673)

功能：本PR的目标是提高attention模块的计算性能。
为了减少框架层对op的调度开销，本PR通过在C++层手动实现attention模块，对外提供attention 大op；
为了减少防存开销，本PR采取了两种优化方法：
（1）在q,k,v计算时通过共享输入X，将该处的gemm，transpose和bias add从三次调用减少为一次；
（2）使用kernel融合优化技术，在不同cuda kernel之间通过寄存器传输数据；

…dle#36454)

* Add new API tensordot
cherry-pick PaddlePaddle#36273

…#36522) (PaddlePaddle#36671)

Refine comments for GradScaler state_dict.

…ePaddle#36672)

Add fp16 kernel for clip_op.

…addlePaddle#35757) (PaddlePaddle#36551)

Add paddle.nn.functional.sparse_attention API

    本个PR主要将sparse_attention功能在python层进行了一层封装，OP的主体代码见：#PR35676

    此外，对于封装的python 接口，增加了相应的单测。

cherry-pick prs

PaddlePaddle#36568
fix fc fuse compat problem

PaddlePaddle#36610
support lite xpu choose device id

PaddlePaddle#36010
update lite branch

PaddlePaddle#36628
add file exists check

…e#36371) (PaddlePaddle#36654)

…e#36549) (PaddlePaddle#36655)

* Fix grid sampler

* Fix code format