Merge pull request !7665 from lyvette/windows

Merge pull request !7728 from pengyongrong/stack

Merge pull request !7712 from yeyunpeng2020/primitive

Merge pull request !7707 from cjh9368/rewrite_aware_quant

Merge pull request !7620 from zhaozhenlong/lite/op/non_max_suppression

Merge pull request !7632 from fuzhiye/tmp

Merge pull request !7627 from 王东旭/opencl_conv_remove_codegen

Merge pull request !7192 from zhengjun10/transformer

Merge pull request !7635 from zhanghaibo/master

2.free reduandent memory in fp16 kernels

Merge pull request !7621 from liuwenhao/master

Merge pull request !7577 from hangq/primitive

*model_buf, size_t size, lite::Context *context) interface

Merge pull request !7579 from gongdaguo/fix_pool

Merge pull request !7467 from wangshaocong/bugfix_master

Merge pull request !7457 from zhanyuan/tmp

Merge pull request !7466 from wangshaocong/bugfix_master_v2

Merge pull request !7501 from chenjianping/lite_windows

Merge pull request !7418 from EricZ/remove_adapter_md

Merge pull request !7327 from 徐安越/master

Brute force fix?

Added missing import

Added removal of new cache functionality

Fix batch

Remove crc and cache optimizer

Remove affine

Remove cut mix batch

Added securec.h

Removed unused crc import

Fixed minor arm64 compile issues for tensor ops

Removed excess code from de_tensor

Added vision import

remove random device ops

Removed commented line

Added extra layer to switch adapter logic

Added securec

Added float fix

Merge pull request !7436 from lixian/master

Merge pull request !7405 from gongdaguo/onnx_fix

Merge pull request !7355 from fuzhiye/tmp