1. add argmax  op

2. add transpose template function

fix single-threaded compilation in CI

Since input tensor is modified to NHWC, which is not compatible to cast
kernel in NCHW layout, type_layout_cast_pass will insert a layout
instruction to transform data. Use NHWC cast kernel to avoid cast layout

fix(quant): fix quant compute error

while filter weight is quanted elementwise, use maximum of
weight_scale instead of mininum to keep quant weight in range of int8

for multiple-use args, only cast once

move sth from device_info to backends/mlu

[MLU] add mlu's missing "APPEND ${fake_kernels_src_list}" in lite.cmake

           testcase

* Optimize matmul for size(x_dims)=2  size(y_dims)>2

* [NPU] add shape bridge

move shape arm kernel to host

* enhance compare arm kernel

* [NPU] add gather op bridge

* enable reshape arm ut

* [NPU] add lookup_table bridge

lite cuda support exec multi-stream

* [LITE][OPENCL][Image] lazy init cl::event , test=develop

* [LITE][OPENCL][Image] lazy init cl::event , test=develop