Cherry-pick PR 38406, fix accumulator bug when multiple inplace OPs are executed continuously  (#38406) (#38830)

Cherry pick PR #38406

Co-authored-by: Nfengkuangxiaxia <fengkuangxiaxia21@gmail.com>

* multithread_memory_optimize

* add mish trt plugin, compile & install success, run error. test=develop

* modify code of mish plugin

* upgrade mish trt plugin

* modify code according to review

* add TRT_NOEXCEPT for mish trt plugin

* add unittest for mish trt plugin

* remove unnecessary check of mish in op_teller.cc

* fix some problem of trt8

* add check and modify unittest while converting mish to trt plugin
Co-authored-by: Ndengkaipeng <dengkaipeng@baidu.com>

* add mish operator and api

* remove redundant code and modify grad_atol of mish unittest

* modify mish code to be consistent with other activation implementation

* modify comment of mish

Cherry-pick PR #38014

* [Cherry-pick] Add the forward QR operator

* Add a check for multiplex op

* Improve multiplex based on reviews

* optimize backward (#37055)

* update

* update

* update

* modify code style

Co-authored-by: Nfeng_shuai <fengshuai03@baidu.com>

[CHERRY-PICK] Fix for matmul_v2 oneDNN op broadcasting when inputs dims have different lengths (#38733)

* fix for matmul_v2 broadcasting

* resolved conflicts

本PR修复了以下两个tensor切片索引bug：

1.修复Tensor索引赋值调用set_value op出现的显存泄露问题，该问题主要是由Inplace策略的使用不当导致，本PR中已完成修复。

2.修复使用多个None类型索引时结果维度异常的问题

* support data_format='NHWC' for prelu channel mode (#37019)

* support data_format='NHWC' for prelu channel mode

* fix prelu weight shape for NHWC of static mode (#38310)

Add python interface of subgraph: 1. all_sub_graphs() 2. get_sub_graph(idx)

优化layer.to实现逻辑，相关pr：
Remove additional warnning in layer.to ( #36700)
Refine param conversion logic in layer.to ( #36862)
Fix Layer.to() of device bug ( #37156)

* fix_ele_convert: IElementWiseLayer can broadcast

* fix_ele_convert

fix: when ceil_model==true && Padding_algo!=SAME, (x-size)/stride != int, this convert is wrong (#37929) (#38033)
Co-authored-by: Nfeng_shuai <fengshuai03@baidu.com>

cherry-pick #37536

修复pass_desc.proto在编译时产生依赖问题。

slice_grad op在选择kernel过程中出现错误，问题原因是在获取use_mkldnn属性时，map中未找到该键值，所以抛出out_of_range异常
本PR在map获取use_mkldnn属性数据前增加了是否存在该键值的判断逻辑，从而避免出现上述异常

cherry-pick #36714

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* [heterps]bug fix for _run_from_dataset

* fix heter_server.cc

* fix launch_utils.py

* fix heter_section_worker.cc

* fix. test=develop

* fix. test=develop

目前的fused_attention_op不支持attn_mask=None的输入，本PR对此进行了补充，并补充了相应的单测逻辑。

Cherry-pick PR 37420, fix inplace bug when the first grad_var(loss_grad) is inplace var (#37420) (#37488)

fix inplace bug，Cherry pick PR #37420

Add support for bias is none for fused_attention op.