Merge pull request !4009 from caojian05/ms_master_dev3

Merge pull request !3983 from huanghui/reduce-min-fission-pass

Merge pull request !3992 from hanjun996/master

Merge pull request !4000 from zhangbuxue/improve_bool_for_tensor

Merge pull request !3985 from chenzhongming/doc

Merge pull request !3966 from lichen/add_dropout_distributed_op

Merge pull request !3989 from lianliguang/test-merge

Merge pull request !3857 from wanghua/master

Merge pull request !3969 from cjh9368/fix_benchmark_inf_bug

Merge pull request !3720 from thlinh/dev_July29_remove_redundant_minmax

Merge pull request !3986 from fangzehua/recipro

Merge pull request !3990 from ZhangQinghua/master1

Merge pull request !3938 from liubuyu/master

Merge pull request !3987 from zhoufeng/graph-compile-performace

Merge pull request !3692 from peixu_ren/custom_bijector

Merge pull request !3972 from kisnwang/update-dnnl-to-v1.5

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

Merge pull request !3807 from 徐永飞/master

Merge pull request !3949 from meixiaowei/master

Merge pull request !3935 from ZhangQinghua/master

Merge pull request !3984 from hangq/master

Merge pull request !3916 from xychow/remove-loss-scale-range-check

Merge pull request !3914 from xiefangqi/md_fix_numpyslice_e2e

Merge pull request !3941 from panfengfeng/fix_gpu_scripts_resnet101

Merge pull request !3946 from YuJianfeng/master

Merge pull request !3828 from 章一智/md-lite