* remove autograd/grad api

* import grad, no_grad set_grad_enable from autograd module

* modify import no_grad_ as no_grad

add fill_ backward

* fix github name

* fix CI error

* fix review and CI error

* fix inf,nan error and modify unittest samples

* add unittest samples

* add unittest samples

* fix unittest error

* test=document_fix

* test=document_fix

* modify doc and add unittest samples

* fix error newline in constant

* modify doc after mentor review

* modify __all__ and doc

* modify doc

* add test_cumprod_op

* Revert "add test_cumprod_op"

This reverts commit c96cf6dff5d09ae7d8cc72c1e8ae4369a153aa19.

* recommit

* add error message

* test input(x) initialize

* test use cpu

* update test code

* add test type

* add test case

* solve ci problem

* add complex case test

* add complex case test

* fix review problem

* fix conflict

* fix some docs

* change test case

* change test case

* fix review problems again

* fix docs

* fix inclusivescan bug

* Add SVD Op and it's GPU and CPU kernel

* Remove CUDAPlace in test_svd_op, make the test available in CPU package

* modfity the file

* fix windows bug/ fix ROCM / fix test timeout

* for pass the CIs

* improve error report

* for code review

* some modification to test_svd_op

* change python code style

* expose the svd interface for document

[Background]
Expansion in code size can be irreversible in the long run, leading to huge release packages which
not only hampers user experience but also exceeds a hard limit of pypi.

In such, NV_FATBIN section takes up 86% of the compiled dylib size, owing to the vast number of GPU
arches supported.

This PR aims to prune this NV_FATBIN.

[Solution]
In the new release strategy, two types of whl packages will be involved:

Cubin PIP package:
PIP package maintains a smaller window for GPU arches support, containing
sm_60, sm_70, sm_75, sm_80 cubins, covering Pascal - Ampere arches

JIT release package:
This is a backup for Cubin PIP package, containing compute_35, compute_50, compute_60,
compute_70, compute_75, compute_80, with best performance and GPU arches coverage.

However, it takes around 10 min to install due to the JIT compilation.

[How to use]
The new release strategy is disabled by default.
To compile for Cubin PIP package, add this to cmake: -DCUBIN_RELEASE_PIP
To compile for JIT release package, add this to cmake: -DJIT_RELEASE_WHL

* Add function to disable paddle signal handler

Paddle used google::InstallFaultSignalHandler to handle selected system signals,
mainly for debugging and bug report purposes.

However, this can be conflicted with other python packages whoever captures similar signals.
Such python package involves tvm and more

To resolve this issue, we support a function to disable signal handler

* Remove signal test from WIN32 platform

* Remove redundant return from disable_signal_handler() function

* Add detailed messages to en_doc

* add unique_consecutive_op

* add unique_consecutive_op

* add unique_consecutive_op

* add unique_consecutive_op

* add unique_consecutive_op

* add unique_consecutive_op

* add unique_consecutive_op

* add unique_consecutive_op

* remove unity build

* add unique_consecutive op

* add unique_consecutive op

* add enable static

* add noqa

* add space line

* add default case.

* add comma

* add space line

* modify unique_consecutive unittest

* optimize ut coverage

* rebase develop

* improve coverage

* update en docs

* update en docs

* update en docs

* update en docs

* update en docs

* update en doc

* OP dot: refactor CPU kernels and get better loop performance.

* Minor fix on code format.

* Fixed minor errors.

* Add new API: einsum

* Update the Einsum unit test.

One case failed with matmul_v2, where the dtype is int64:

a = np.arange(2 * 3 * 1).reshape(2, 3, 1)
b = np.arange(1)
paddle.einsum("...i, ...i", a, b)

* Test cases in test_einsum test floating point dtypes only.

As of now Paddle only supports float/double dtypes in matmul, which is
one of building blocks of this Einsum implementation. We decide not to
test einsum against other dtypes.

* Polish format.

* More formatting.

* Format...

* Einsum: improve test coverage.

* Einsum: bug fixes and more testcases for testing error messages

* Einsum: fix format..

* Einsum: fixed typo and format.

* Einsum: format again...

* Einsum: applied suggested changes.

* Einsum API: improve API documentation.

* Einsum API: apply suggested changes.

* Einsum API: Add dygraph only note.

* Einsum API: Add dygraph only note.

* Einsum API: fixed unittest.

* add cuda event and stream api

* add cuda event and stream api

* add get_current_stream api

* add get_current_stream api

* init streams

* modify get_current_stream

* modify get_cuttent_stream

* add synchronize func

* add current_stream doc and test file

* move get_current_stream into CUDA macro

* move CudaEvent into CUDA macro

* move _get_current_stream and _device_synchronize into cuda macro

* modify the macro of cuda stream and event

* add test case for synchronize

* add paddle.devices.cuda module

* event and stream support hip

* add doc for stream and event class

* move cuda stream and event into single pybind

* add cuda_streams_py.cc to cmakelist

* add _device_synchronize and _get_current_stream to core module

* add test case for cudastream and cudaevent

* move __all__ in streams.py

* fix test fail

* add cuda to devices __all__

* fix current_stream doc writing error

* move devices to device direction, and merge device.py into __init__.py

* add required:gpu to sample codes

* remove cuda direction from device/__init__.py

* new api diagonal, test=develop

* add new api diagonal, test=develop

* new api diagonal, test=develop

* add new api paddle.diagonal, test=develop

* use framework::stride replace ComputeDimStride

* replace cudaMalloc/cudaMemcpy by TensorFormVector in cudaKernel and cudaGradKernel

* perfect funciton: when attr(offset) is exceed attr(axis1) or attr(axis2), set the diagonal dim is 0

* fix RP-Mac-CI bug: replace framework::stride() by ComputDimStride.

* perfect code-block

* perfect code of python API diagonal

* api supports dtype of float16 and bool

* api supports dtype of float16 and bool

* modify unittest code

* modify unittest code

* perfect dtype describe

* perfect code-block

* add atan2_op

* fix

* new api trunc, test=develop

* Add digamma_op and unittest

* add digamma_op api

* remove special DigammaCudaKernel and correct some docs

* remove unused headers

* fix api doc error

* update 2.0 public api in all left files

* reverse device.py all list;
fix some flake8 errors

* add paddle.neg api

* add test for neg

* fix an English gammar error in comment

* add lgamma api

* support api paddle.tensor.neg() and paddle.tensor.lgamma()

* modify test_neg_op.py

* [OPs] Bug fix, fix the segment mean for illegal syncthreads usage. (#32596) (#32610)

* [OPs] Bug fix, fix the segment mean for illegal syncthreads usage.

* remove packages in __all__

* create new public api level paddle.callbacks;paddle.hub;paddle.utils.unique_name
Co-authored-by: NZhong Hui <zhonghui.net@gmail.com>

Co-authored-by: NXiaoguangHu <46782768+XiaoguangHu01@users.noreply.github.com>

* add Hub Module for easy to use pre-trained models.
*   support list, load, help fucntions.
*   support load models by github, gitee, local 
Co-authored-by: NLielinJiang <jianglielin@baidu.com>

* add common dtypes as paddle's dtypes

* import paddle.fluid.core_avx.VarDesc.VarType as paddle.dtype

* [feature] support npu allocator (#30840)

[feature] support npu allocator

* [feature] support npu operator (#30951)

[feature] support npu operator

* [feature] support npu allocator, part 2 (#30972)

* support npu allocator

* add npu device context

* fix some compile problem

* fix some compile problem

* add npu info

* compile ok

* fix include dir

* support naive_best_fit_allocator

* run ut ok, bug failed to exit

* call aclrtResetDevice before exit

* fix aclFinilize

* add system allocatot test

* add selected_gpus in gtest

* add tensor_test for npu

* support npu op, initial commit

* add npu stream

* add elementwise_add_op

* compile ok

* fix typo

* fix elementwise_add_op_npu_test

* support op run

* test can run but failed

* change aclopExecuteV2 to aclopCompileAndExecute

* support parsing ascend rank table file (#31000)

support parsing ascend rank table file

* Fix reshape on GE graph. (#31084)

Fix reshape on GE graph

* add npu kernel for elementwise_sub and elementwise_sub_grad (#30973)

* add npu sub op

* fix typo

* rename test

* fix bug

* fix bug

* add fp16 kernel

* fix typo

* support sub grad op

* support elementwise_sub_grad op
Co-authored-by: Nfrankwhzhang <frankwhzhang@126.com>

* Fix compilation problem (#31100)

Fix compilation problem (#31100)

* fix compile

* fix code stype

* remove const_cast

* support adding correct npu op in pybind.h (#31143)

* support adding correct npu op in pybind.h

* refine code

* [NPU] Support executor with NPU (#31057)

* [NPU] Support executor with NPU

* Fix code according to reviews

* Fix code

* Add unittest for sub op npu

* refactor npu device manager (#31154)

refactor npu device manager (#31154)

* fix selected npus

* fix compile

* fix reading flags from env

* format
Co-authored-by: Nxiayanming <41795079@qq.com>
Co-authored-by: Ngongweibao <weibao.gong@gmail.com>
Co-authored-by: Nfrankwhzhang <frankwhzhang@126.com>
Co-authored-by: Nliym27 <33742067+liym27@users.noreply.github.com>

* add custom init grad for backward function

* add custom init grad for backward function

* handle when the grad_tensor is none

* handle when the grad_tensor is none

* fix the args type error on windows platform

* modify the args order and doc

* format code

* add grad_tensor to xpu

* modify the grad_tensor type check

* add paddle.backward api to support multi tensors gradient compute

* add paddle.backward api to support multi tensors gradient compute

* add paddle.atuograd module and backward api

* change tensor.backward func args

* modify tensor backward api

* remove create_graph intputs args

* add doc and examplex code for backward api

* when have the same tensor, throw error

* modify test Init func args

* modify the execute.Init func args in test files

* add paddle.autograd package in setup.py.in

* modify error msg, remove _run_backward method in class Tensor

* add test cases for backward api

* add view strategy on squeeze,unsqueeze,reshape,flatten

* add squeeze unittest

* add unittests

* use View strategy as name rather than Reuse Allacation

* fix view api doc

* fix format

* use core.ops when input of reshape2 is Tensor

* fix test_cross_entropy_loss error because of reshape2

* fix test_cross_entropy_loss error because of reshape2

* add inplace strategy

* add elementwise_add sub

* let backward op not use inplace

* grad op do not use inplace

* fix memory increase error and add leaf error message

* delete selected_rows

* change op_function

* little change

* solve HandleViewBetweenInputAndOutput

* add unittest and leaf error message

* merge view error

* optimize op_function_generator format and support sum inplace op

* fix format of basic_engine

* fix format for framework

* little change of variable wrapper

* add reshape, squeeze, unsqueeze, scatter api

* add relu elu tanh softmax inplace api

* fix test_squeeze_op unittest

* fix test_relu_op unittest

* fix comment problems

* delete sample code of inplace api

* add reference of grad_pending_nodes in basic_engine

* fix unittest name

* add inplace apis into wlist

* fix error message

* add PADDLE_ENFORCE for set grad op twice

* fix head file error

* test=develop, add model_average and lookahead