Merge pull request !4063 from ZhangQinghua/master1

Merge pull request !4001 from chenzhongming/master

Merge pull request !4099 from zhaizhiqiang/master

Merge pull request !3813 from yanghaoran/master

Merge pull request !4064 from huanghui/reduce-min-fission-pass

Merge pull request !3944 from xiefangqi/md_new_api_CelebA

Merge pull request !4085 from fuzhiye/tmp

Merge pull request !4085 from fuzhiye/tmp

Merge pull request !4082 from peixu_ren/master

Merge pull request !4086 from wandongdong/master

Merge pull request !4098 from zhoufeng/graph-compile-performance-optimize

Merge pull request !4031 from riemann_penn/fix_cell_bprop

Merge pull request !3931 from huangxinjing/synthetic_data

Merge pull request !3882 from JoyLvliang/decoupling-the-interface-of-mallocing-mem

Merge pull request !4091 from hangq/master

Merge pull request !4039 from wangzhe/master

Signed-off-by: zhoufeng <zhoufeng54@huawei.com>

Merge pull request !4093 from sunsuodong/relu6_int8

Merge pull request !4077 from wangchangkai/master

Merge pull request !3951 from songhonglei413/test

Merge pull request !4079 from yankai10/new_merge

Merge pull request !4006 from Simson/enhancement-API

Merge pull request !4062 from yangruoqi713/fix_bug

Merge pull request !4010 from changzherui/mod_ckpt_param

Merge pull request !4072 from hewei/fixbug_tensor_equal

Merge pull request !4050 from fangzehua/unsortedgrad

Merge pull request !4080 from andy_wangrui/master

Merge pull request !4056 from liubuyu/bug_fix

Merge pull request !4026 from 34bunny/CMake-debug-mode-fix

Merge pull request !3831 from danishnxt/GPU_One

Merge pull request !3961 from lixian/master

Merge pull request !4040 from sunsuodong/strided_slice_int8