* add multiclass_nms3 converter

* fix bug that some op has no op_role attr

* add mkldnn support for new executor

* fit for mkldnn data_transfer

* fit for mkldnn data_transfer

* 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 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

* op:transpose_op supports bool type

* Keep strided_slice op behavior consistent with slice op when starts input is less than -rank

* Using DistConfig in inference

* no

* maintain old profiler

* exclude new python record events for old profiler

* maintain old profiler

* maintain

* maintain old profiler

* maintain

* fix cmakes

* add flatten2,reshape2,squueze2_trt_fuse_pass  test cast

* add flatten2,reshape2,squueze2_trt_fuse_pass  test cast

* add flatten2,reshape2,squueze2_trt_fuse_pass  test cast

* mv yolov3_loss op to phi

* fix as review

* update operator.h

* fix load bug and add distributed strategy from pslib

* add unittest

* use cvm config

* trainer and worker config

* add unittest

* add unittest

* add test

* code style

* fix dependency of fused optimizer

* add ut

* no

* fix bugs

* fix doc according to review

* fix api doc format

* fix api doc according to review

* fix bug and add unit test

* fix record event bug

* optimize chrome tracing display

* fix bug

* add comment

* add unit test

* fix a bug

* fix

* fix

* fix format

* [KP] fix bug in phi kp

* delete useless comment

* update

* update

* choose the xpu kp kernel in phi

* for 1st time interface combine.

* modification with kernel factory

* first auto_tune version.

* first version.

* basic version

* add warm up step.

* a debug version.

* optimize the functionality of class auto_tuner.

* add some quotes for optimized auto_tuner class.

* add some quotes for optimized auto_tuner class.

* add namespace.

* modification according to the advices

* replace fluid header with phi header.

* replace fluid header with phi header.

* Restrict compilation conditions of optimized topk kernel

* fix

* add weight unfold pass and handle trt fc op

* update

* add kernel

* update

* update

* support view strategy in eager_fluid state

* little change

* little change

* optimize unittest

* fix

* Supported Complex2Real Conversion for Eager Dygraph

* Supported Complex2Real Conversion for Eager Dygraph

* Enabled complex type promotion test for matmul_v2

* pylayer, test=develop

* Fix CI issues

* Support initializing specific grad tensors to zero for selected operators

* finish forward, test=develop

* create grad node finish, test=develop

* Merged adj_edges_ with GradSlotMeta

* Fixed monir issue

* backward finish, start dbg, test=develop

* Adjusted num runs

* Recovered Eager performance tests configurations

* Recovered Eager performance tests configurations

* finish, test=develop

* polish, test=develop

* polish, test=develop

* refine, test=develop

* eager, test=develop

* Adjusted performance tests configurations

* Fixed Minor Issues with performance tests

* [Phi] Fix macro name typo

* support set_materialize_grads, test=develop

* suppotr mark_non_differentiable, test=develop

* support once_differentiable, test=develop

* refine, test=develop

* refine, test=develop

* Moved out Edge from GradSlotMeta

* Fixed issues from merge

* Fixed typo

* Addressed review comments

* Fixed merge issues

* Fixed minor issues

* Fixed minor issue

* refine, test=develop

* refine, test=develop

* refine, test=develop

* Fixed major issues and enabled auto_prune test cases

* Fixed issues from merge

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* support inplace for pylayer
Co-authored-by: Njim19930609 <jim19930609@gmail.com>
Co-authored-by: NWang Huan <wanghuan29@baidu.com>
Co-authored-by: NAurelius84 <zhangliujie@baidu.com>

* move rnn kernel to phi

* move infershape of rnn to phi

* fix HIP bug

* rename function

* fix HIP bug

* fix hip bug

* add random routing op

add _random_routing api in utils

add random routing ut

* # This is a combination of 10 commits.
# The first commit's message is:
add expert count op

add ut for expert_count

# This is the 2nd commit message:

update UT only for cuda

# This is the 3rd commit message:

fix for rocm

# This is the 4th commit message:

update ut

# This is the 5th commit message:

add moe module

# This is the 6th commit message:

add expert count op

add ut for expert_count

# This is the 7th commit message:

update UT only for cuda

# This is the 8th commit message:

update ut

# This is the 9th commit message:

add moe module

# This is the 10th commit message:

make expert count private

* add assign pos op

* fix upper num name

* add api _assign pos

* add ut for assign pos op

* update date

* add op about moe gate

update utils

add limit by capacity op

add ut for limit_by_capacity

add ut for prune_gate_by_capacity

add ut for limit_by_capacity

add ut for prune_gate_by_capacity

* fix for win

* fix bugs in test_limit_by_capacity_op

* update ut

* update for test (timeout)

* fix ut

* update

* update(fix) ut for win

* moe apis in incubate

* # This is a combination of 10 commits.
# The first commit's message is:
add expert count op

add ut for expert_count

# This is the 2nd commit message:

update UT only for cuda

# This is the 3rd commit message:

fix for rocm

# This is the 4th commit message:

update ut

# This is the 5th commit message:

add moe module

# This is the 6th commit message:

add expert count op

add ut for expert_count

# This is the 7th commit message:

update UT only for cuda

# This is the 8th commit message:

update ut

# This is the 9th commit message:

add moe module

# This is the 10th commit message:

make expert count private

* add assign pos op

* fix upper num name

* add api _assign pos

* add ut for assign pos op

* update date

* fix for win

* update for test (timeout)

* fix ut

* update

* fix ut for number count

* add apis and utils

* add gate apis

* add moe and grad clip apis

* update moe apis

* add ops for moe gate

* fix

* update for base moe layer api

* add random routing op

add _random_routing api in utils

add random routing ut

* fix for dygraph

* update with ranodm routing

* update

* fix ut for limit by capacity

* update

* update limit by capacity for easily to switch to single thread mode

* update api docs
Co-authored-by: Nhlygit66666 <2570058140@qq.com>

* fix uncontrolled randomness of op

* fix bugs

Revert "Revert "[Phi] Move elementwise_floordiv and elementwise_pow to phi (#40993)" (#41065)" (#41110)

This reverts commit 3a6f1135.

Add new APIs for GPU memory monitoring (max_memory_allocated, max_memory_reserved, memory_allocated, memory_reserved) (#38657)

* Add new API memory_reserved

* Add memory_allocated, max_memory_reserved and max_memory_allocater

* Fix CI error

* Fix CI error

* Enhance UT

* Add FLAGS_memory_stats_opt

* Add STATS macro functions

* Add StatAllocator

* Fix CI errors

* Add UT

* Fix CI errors

This reverts commit 054fc997.

This reverts commit 05f3d48e.

* [Refactor] refactored eager_gen.py PR #2

* [DoubleGrad PR #1] Decoupled code generation logics for Dygraph ForwardFunctions and GradNodes

* Fixed minor issue

* Adjusted logics of GenerateNodeCreationCodes and GenerateForwardDefinition

* Fixed issues

* Supported higher-order grad node generation

* [DoubleGrad PR #4] Supported higher-order GradNode generation

* Fixed yaml typo

* [Yaml] Fix topk yaml compilation on Windows

* fix make_shared

* fix conflict

* add bilinear interpolate v2 to xpu list and unitteset, *test=kunlun

* Delete ps_usr_print_log

* Delete ps_usr_print_log

* Delete xpu_op_test