* fix non-sse non-neon weight pack

* update riscv qemu

* c906 test on qemu

* fix qemu aarch64

* simplify binaryop

* less gpu test

* update binaryop broadcast doc

* do not test atan2 zero

* write shape, fuse sam image encoder attention

* set more dynamic shape as static

* less warning for constant tensor node

fuse F.scaled_dot_product_attention

Bumps [pypa/cibuildwheel](https://github.com/pypa/cibuildwheel) from 2.12.3 to 2.13.0.
- [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/v2.12.3...v2.13.0)

---
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>

* fix overwrite builtin layer destruction

* make modelbin class copyable

* test++

* pnnx fuse transformer clip attention

* skip fuse mha for 1.8

* select one method other than forward

* pnnx fuse diffusers attentionblock

* force global cpu info initialization

* sanitize zero nT

* pnnx convert sdpa

* pnnx fuse diffuser attention2

* Update benchmark information about Hyper-V Linux Guest with GPU-PV enabled (Intel Core i7-11800H, NVIDIA GeForce RTX 3070 Laptop GPU).

* Update benchmark information about Hyper-V Linux Guest with GPU-PV enabled (Intel Core i7-7700K, NVIDIA GeForce GTX 1050 Ti).

* fix vulkan validation error, prefer VK_KHR_buffer_device_address over VK_EXT_buffer_device_address

* enable validation extension features

* workaround moltenvk error on spec const composite op

* workaround moltenvk crying on binding image with memory offset

* pnnx convert torch std

* fix multiple fuse pass on torch 2.0

* fuse vit pytorch mha pattern

* fix aarch64 build without fp16 conversion intrinsics

* vfpv4 always implies neon

support yolov8s ncnn