Merge pull request !2844 from kisnwang/optimize-cpu-mem

Merge pull request !2333 from xychow/remove-addnclass-in-opt

Merge pull request !2819 from ghzl/add_group_param_order

Merge pull request !2840 from YuJianfeng/master

Merge pull request !2841 from Kang/master

Merge pull request !2823 from xutianchun/fix_doc

Merge pull request !2820 from zyli2020/refine_mem_swap

Merge pull request !2818 from ddwolf/add_primitive_op_to_test_lamb

Merge pull request !2828 from zhangbuxue/fix_tensor_index_when_there_is_only_one_element_in_tuple

Merge pull request !2817 from kisnwang/optimize-sub-graph-memcpy

Merge pull request !2814 from jiangjinsheng/issue_fix4

Merge pull request !2810 from zhangzheng/softsign

Merge pull request !2832 from Peilin/invalid-type-bucket-check

Merge pull request !2831 from thlinh/dev_Jul02_rework_arithmetic_simplify

Fix memcpy calls. Add ut tests for arithmetic_simplify. Split long arithmetic_simplify.h to arithmetic_simplify.cc

Code checking

Merge pull request !2738 from islam_amin/object_ops_ut_cc

Merge pull request !2821 from yanghaitao/yht_fix_device_queu_profiling

Merge pull request !2822 from hewei/fix_bool_tensor

Problem: Create 'uint8' type tensor from float data raise RuntimeError;
Fix:  Use same internal data type for 'uint8' and 'bool' tensor and allows
convert float data to bool tensor.

Merge pull request !2803 from shenwei41/sw_beifen

Merge pull request !2013 from liuwenhao/pr_1949

Merge pull request !2478 from BowenK/python_pass_clean

Merge pull request !2804 from liuxiao93/ops-for-VM

Merge pull request !2565 from ddwolf/add_case_for_precisoin_of_bert

Merge pull request !2806 from zhoufeng/process-empty-graph