* add clear_gradients for Optimizer, add api description, test=develop

* fix optest for optimizer's clear_gradient interface, test=develop

* add sample code, test=develop

* polish sample code, test=develop

* update layers, test=develop

* update layers for resnet, test=develop

* fix is_test attr, test=develop

* update cycle_gan, test=develop

* update reinforcement_learning, test=develop

* update ocr, test=develop

* fix bug, test=develop

* Polish the PADDLE_ENFORCE in fusion_group pass related codes.
test=develop

* Correct the unittest because of the change relu_grad's formula.
test=develop

tweak the interface of cache_third_party function - expose the SOURCE_DIR for each external library (#21899)

* update PRelu arg
test=develop

* fix unittests
test=develop

* fix element mode alpha shape
test=develop

* split channel_or_input_shape arg
test=develop

* fix unittest
test=develop

* add support for nested profiling event and printing in different level

* add example for reshape op

stack's wait cost a lot of cpu time, use cuda kernel to do memory copy
will reduce cpu time.
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* modify fc to linear in sample code, test=develop

* remove FC, test=develop

* remove warnings, test=develop

* drop fluid/imperative/README.md , test=develop

* change fc to linear, test=develop

* polish code style, test=develop

Add coverage data processing tools.

1. Add a new input named batch_roi_nums for prroi_pool_op. batch_roi_nums includes the number of roi for each image in batch when rois is Tensor. This information is saved in rois's lod when rois is LoDTensor.
2. add grad check to prroi_pool_op and solve unnormal X grad diff in CPU.

* fix sample code, test=develop

* polish code style, test=develop

* modify test framework, test=develop

* fix test_fsp_op, test=develop

* up, test=develop

* up, test=develop

* up, test=develop

* add check for threshold, test=develop

* up, test=develop

* up, test=develop

* up, test=develop

* up, test=develop

* up, test=develop

* up, test=develop

* up, test=develop

* update, test=develop

* update, test=develop

* update, test=develop

test=develop

enhanced ops: acos, asin, brelu, ceil, cos, elu, floor, hard_shrink,
hard_sigmoid, hard_swish, relu6, round, rsqrt, sin, softshrink,
cos_sim, cross_entropy, im2squence, kldiv_loss

* enhance load interface; test=develop

* add uni test and add comment; test=develop

* fix converage; test=develop

* use path.joint replace "/"; test=develop

* windows debug; test=develop

* fix window unitest error; test=develop

* fix commet error; test=develop

* add model shuffix check; test=develop

* fix example error; test=develop

test=develop

* add TRT support for instance_norm op

windows conv_fusion failed as no kernel， explicit declare lambda
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* support dygraph structured name; test=develop

* add load static param unitest and fix save load; test=develop

* fix varBase import error; test=develop

* fix unitest error; test=develop

* add comment for parameter; test=develop

* fix uni test error; test=develop

* change parallel se-resnet; test=develop

* fix dygraph se resnext parallel test; test=develop

* remove useless code; test=develop

* remove useless code; test=develop

* set esp as 1e-6 to solve elu unitest fail,test=develop

* add special way to add distribute vars， Update Pyramid hash op

* fix Variable's gradient api in framework.py, test=develop

* remove namescope, test=develop

* Fix the global_step & continuous applying error in EMA

test=develop

* Fix for step 0 & add unit test, test=develop