* split arm82 source, wip

* check compiler arm82 only for arm 64bit target

* drop arm82 registery

* strict check compiler support arm82

* discard riscv innerproduct weight
* drop riscv conv convdw weight
* drop riscv deconv deconvdw weight

* discard weight memory for x86 and vulkan
* drop arm innerproduct weight
* drop arm convolution weight
* drop arm convolutiondepthwise weight
* drop x86 vulkan deconvolution deconvolutiondepthwise weight
* drop arm deconvolution deconvolutiondepthwise weight
* arm neon assembly optimization for innerproduct pack4

* riscv general optimization for convolution sgemm and winograd pack1
* riscv general optimization for innerproduct
* riscv general optimization for convdw3x3

* mips optimization for convolution sgemm

* mips optimization for general convolution int8 gemm

* mips optmization for convolution winograd pack1

* preload magic

Bumps [pypa/cibuildwheel](https://github.com/pypa/cibuildwheel) from 2.5.0 to 2.6.0.
- [Release notes](https://github.com/pypa/cibuildwheel/releases)
- [Changelog](https://github.com/pypa/cibuildwheel/blob/main/docs/changelog.md)
- [Commits](https://github.com/pypa/cibuildwheel/compare/2.5.0...2.6.0)

---
updated-dependencies:
- dependency-name: pypa/cibuildwheel
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>

Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

* optimize x86 winograd input transform transpose

* x86 sse2/avx optimization for convolution winograd23/43 pack1

* chain dict output

* recognize sub_

* feat(src/layer): add vision_transformer benchmark and relative layer

* refactor(testutil.h): add para for RandomMat

* RVV: follow intrinsic doc, vfredusum -> vfredsum

* C906: change toolchains for vfredusum

* RVV: test compiler for vfredusum_vs_*

* bitwise and compare op

* masked_fill and gather test