Merge pull request !29443 from lyqlola/master

Merge pull request !29445 from zhuzhongrui/pub_master4

Merge pull request !28935 from lingyunli63/fix_NoOpNode

Merge pull request !28728 from fangzehua/mkl_tp

Merge pull request !28546 from kisnwang/add-cpu-adafactor

Merge pull request !28155 from lingyunli63/rm_dynamic_reshape

Merge pull request !28726 from caifubi/master-pynative-mindrt-dynamic-shape

Merge pull request !27898 from yuchaojie/op_select

Merge pull request !28113 from liubuyu/bug_fix

Merge pull request !28197 from kisnwang/add-cpu-adafactor

Merge pull request !28077 from fangzehua/dynamic_batch_mindir

Merge pull request !27975 from DeshiChen/1220_internaloutput

the output address of internal output nop node is required by the next KernelGraph,
so we cannot replace it in graphkernel optimization.

The internal output nodes of KernelGraph are stored in a `map`
for memory reuse between KernelGraphs.

However, the nodes may be changed in graphkernel optimization.
the passes of graphkernel are base on `FuncGraph`, instead of `KernelGraph`,
so we cannot use the `ReplaceInternalOutput` interface.

This commit, we add an attribute for these nodes and skip them in graphkernel's passes.

Merge pull request !27066 from kisnwang/enable-set-swap-node

Merge pull request !26976 from yuchaojie/op_select

Merge pull request !26370 from yuchaojie/ir_fusion4

Merge pull request !26006 from yao_yf/add_transformer_slice_activation_config

slice recompute activation

Merge pull request !26310 from linqingke/aicpu

Merge pull request !25592 from wangnan39/reshape_support_tensor