Merge pull request !2546 from liuwenhao/r0.5

Merge pull request !2539 from geekun/yjk_r0.5

The summary operator will be optimized when it return the origin value
in constant folding scene. So I return a None value to avoid this.

Merge pull request !2412 from jiangjinsheng/issue_fix4

Merge pull request !2417 from liuwenhao/master

- It provides a unified style to express graph and kernel for user.
- It provides a unified IR to represent graph and kernel for developer.
- It breaks the boundary between graph and kernel.
- It provides more opportunities to do compile optimization.

Merge pull request !2356 from zhaozhenlong/fix-issues-broadcastto-same-shape

Merge pull request !2381 from liangzelang/dev

Merge pull request !2402 from liuwenhao/master

Merge pull request !2391 from BowenK/op_elim

make unsupported shape error info explicit

Merge pull request !2406 from jiangjinsheng/issue_fix4

Merge pull request !2060 from amongo/SupportPynativeSetItem

Merge pull request !2365 from wangqiuliang/fix-parameter-cast-issue

support for tensor setitem

add support for tensor assgin

Merge pull request !2347 from riemann_penn/support_multi_param_for_tuple_grad

Merge pull request !2259 from pkuliuliu/master

Merge pull request !2359 from yihuaijie/master

Merge pull request !2351 from liuwenhao/master

Merge pull request !2264 from liuxiao/Adapt-ApplyProximalAdagrad-operator-for-vm

Merge pull request !2328 from jiangjinsheng/issue_fix4

Merge pull request !2315 from yanzhenxiang2020/add_pack_open

Merge pull request !2268 from yanzhenxiang2020/open_ScatterMax_bprop

Merge pull request !2331 from Simson/push-to-opensource

Merge pull request !2161 from vlne-v1/I1IZV3-quant-infer

Merge pull request !2282 from chenzhongming/master

Merge pull request !2275 from Xiaoda/4-fix-the-embeddinglookup-errors