修复《jit.save在Mac系统上保存的模型，在Linux平台上无法对模型进行重训练》的问题。

原始PR： #34027

* Save all the information of 'ParamBase' in 'Layer'. (#33500)

* Save all the information of 'ParamBase' in 'Layer'.

* edit unittest

* delete the function of saving layer object. (#33697)

* delete the function of saving layer object.

* edit doc of paddle.save/load and polish error message

* update xpu cmake for kunlun (#33328)

* xpu support amp (#33809)

* fix bug DLTP-31078 (#33877)

* update xpu cmake (#33906)

* [xpu] add dropout & amp ops in xpu place (#33891)
Co-authored-by: NTTerror <tangzhiyi11@users.noreply.github.com>

[Cherry-pick] Up cxx11 check to cxx14 #34034

[oneDNN] Fix to #33282 , added support of X input broadcasting to oneDNN elementwise ops (#33549) (#33845)

* - fix to #33282

* - Increased threshold for elementwise_mul_bf16 grad

* -disabled faulty UT

* - fix to approval

* del python2 code

* cherry-pick 33556

Bug fixes， [Cherry-pick][Dy2Stat] Fix unique_name in create_static_variable_gast_node (#33963)

将关于approve相关的修改cherry-pick到2.1分支

cherry-pick #33818

[Dy2Stat] Support Python3 type hint (#33745)

fix bug when the cuda kernel config exceeds dims max

cherry-pick Specify gast version in requirements.txt

解决windows cmd和powershell显示乱码的问题

* [XPU] Update cmake options for xpu. (#33450)

* patch in inference third_party

cherry-pick #32815

cherry-pick 32784

[Inference Tensorrt] Add attr for trt engine and handle the input seq problem for ernie var len. (#33575) (#33622)

修复Windows CI的一些突发新增的不稳定现象：

    清理服务器后台未正常退出的进程
    支持外部手动修改cache缓存目录

* cherry-pick fix_output_padding_conv

* add repr unittest for conv

* fix gather op and add logsumexp op on kunlun

* update xpu depence

* update tests and fix elementwise_add

* Add raw program meta optimizer (#32597)

* add raw program, test=develop

* add precision unitest for executor all reduce (#33339)

* fix dp (#33297)
Co-authored-by: NYuang Liu <liuyuang@baidu.com>
Co-authored-by: N李季 <2042519524@qq.com>

* [oneDNN] First fix to #33021  (#33174)

* - First fix to #33021

* [oneDNN] Second fix to #33021 (#33471)

* use older download_data function
Co-authored-by: NJacek Czaja <jacek.czaja@intel.com>

* 1, remove layernorm dynamic fp16; 2, let reshape out in dynamic shape (#33535)

* add win_capi_tar in paddle_build.bat

* tar capi lib for publish

* add in gen_fluid_lib func