Update the error message for the margin_ranking_loss

* add cross_entropy to nn/layer and nn/functional, test=develop

* use functional/cross_entropy in layer/CrossEntropy

* use functional/cross_entropy in layer/CrossEntropy, test=develop

* refine load model by extra_var_info

* polish unittest for coverage

* consider the combination of different strategies to work together

* Rename uniform and gaussian APIs

* rename uniform_random API as uniform

* Fixed unittest

* fixed unittest AttributeError

* Fixed unittest

* Add uniform function rather than alias

* remove templatedoc

* solve conflict and fix doc code

* fix doc code

test=document_fix

* fix decay global counter

* remove unused print, test=distp0

* binding tensor method

* binding tensor method

* binding tensor method

* Binding methods for class Tensor and Variable

* Update `paddle.clamp`

rename to `paddle.clip`
add fast path for dygraph mode
remove `out`
rename `input` -> `x`
update doc sample

* Fix leftover `Variable` wording

* Indent doc with spaces

* Remove `:alias` in docs

* Update `enable_imperative` -> `disable_static`

* Remove `imperative`

also trigger CI

* Update tests for better coverage

* Rebase to fix `cosine_similarity`

* Fix `cosine_similarity` some more

* fixed static module

* solve conflict

* Add Distribution base class, Uniform class and Normal class

* release Distribution class and Normal, Uniform class

* Add Doc args explaination

* save distributions.py and process in distribution.py

* delete useless function in test_distribution

* Add NormalNumpy test class

* Add probs in NormalNumpy

* add distribution to paddle init

* Add Distribution base class and name attribute unittest

* Change Doc

* Change Doc

* adjust format

* Fixed Doc Code

* implement probs and change Variable to Tensor

* Add name for all functions and add name unittest

* support int datatype

* Add dynamic mode

* optimize test_distribution static and dygraph

As the title

* add adaptive_avg_pool2d

* add adaptive_avg_pool3d

Remove cudnn conv in Unittest because the cudnn conv, which is assigned to OP owner, hasn't been fixed.

update the test_mean test case

Fix the API of bce loss and add functional API binary_cross_entropy 

* fix static dataloader default pin memory set

* fix related unittests

disable test_dist_fleet_ctr

* Add new API: paddle.prod

test=develop

* verify correctness of dy2static model by using paddle-inference api

* update python doc style

* disable UT temp

* add print test_dist UT elapsed time

* UT elapsed time optimize

* Rename `Input` into `InputSpec`

* remove alias

* add weight_norm, test=develop

(1) Add api paddle.reshape(x, shape, name); 

(2) Use Tensor replaces Variable. test=develop