Merge pull request !3823 from chenzupeng/master-lite

Merge pull request !3881 from 张学同/to_merge

Merge pull request !3890 from ling/conv1x1

Merge pull request !3103 from yihuaijie/dev

Merge pull request !3877 from liuwenhao/master

Merge pull request !3880 from wangqiuliang/fix-gradients-issue-in-pynative

Merge pull request !3875 from zhaozhenlong/lite/issue/reduce_for_loop_size_unsign_compare

Merge pull request !3820 from zhangbuxue/support_dtype_and_shape_as_attr_in_graph_mode

Merge pull request !3866 from fuzhiye/mindspore

Merge pull request !3865 from yankai10/merge_branch

Merge pull request !3868 from 张学同/to_merge

Merge pull request !3816 from zhouyuanshen/master

Merge pull request !3856 from fangzehua/fix_matrix_msg

Merge pull request !3873 from zhongligeng/master

Merge pull request !3473 from vlne-v1/change_base_class_of_parameter

Merge pull request !3799 from ling/conv1x1

if funcgraph or metafuncgraph call it recursively, then anf_node should be used as tracking_id to discriminate the first occurcance and the
recursive occurance.
add anf_node to PrimitiveAbstractClosure hash() to reduce cost of GetEvaluatorFor().

ignore the tracking_id to make cse work.

Merge pull request !3859 from yao_yf/embeddinglookup_parallel_ops_host_device

Merge pull request !3863 from lijiaqi/bug

Merge pull request !3716 from jiangzhiwen/dataset/adapt_eof

Merge pull request !3774 from JoyLvliang/sync-device-data-type-to-return-tensor

Merge pull request !3822 from hewei/decouple_debug_trace

Merge pull request !3736 from chenzhongming/master

Merge pull request !3642 from lingyunli63/use_full_akg_to_support_gpu