* [NPU] add AMP O1 support

* [NPU] fix NOTE and warnings

* refine amp.decorate code example

* refine code

* refine bf16 amp-o1 logic

* refine amp GLOG

* refine unittest

* refine unittest

* fix pylayer problem with amp

* add ut

* refine code

* fix 'is with a literal'

* fix typo

* support dtype param for auto_cast

* add amp_dtype for tracer

* add unsupported bf16 list

* support bf16 amp for O2

* refine python interface for bfloat16

* refine code

* refine code

* refine unittest

* refine code

* refine code

* add bf16 o1

* refine code by comment

* add gradient accumulator

* add recompute

* check amp.decorate and DataParallel

* refine coverage

* fix layer dtype

* refine code

* add bn_1d_2d_3d for fp16 decorate

* add unittest

* Fix scatter_op fp16 perf problem.

* Add scatter into black list.

* Add scatter into black list for dygraph.

* fix bug

* refine code

* refine code

* refine code

* amp.decorate optimizers set to None is ok

* refine unittest

* add unittest and refine example code

* refine unittest

* run dy2stat pure fp16 in Linear model

* no use self._pure_fp16_inputs

* add test and fix Adam error in dy2stat pure fp16 training

* use paddle.optimizer.Adam

* run test in gpu

* change test time for CI

* enlarge atol for test_resnet_pure_fp16

* refine code and enlarge atol

* make custom_white_list and custom_black_list take effect for AMP and pure fp16

* check tracer is not None

* use default atol

* change filter_size

* change atol and add some NOTE

* refine layer to

* delete comment

* refine logic

* refine code

* refine pure_fp16_init

* refine comment

* [hapi] support dygrapg amp O2

* fix problem of static pure fp16 in hapi

* fix bug

* fix format

* fix ut

* follow comments

* update ut

* update amp save/load

* fix ut

* refine code format

* add attr is_distributed

* refine code

* refine black/white list for pure fp16

* refine amp level

* fix typo

* update tracer._amp_level

* split minimize() to step() + update()

* add unscale and step for grad_scaler

* add unittest

* refine code in minimize

* delete step in loss_scaler

* fix example bug

* refine comment

* refine unittest

* add unittest

* add pure fp16 major function in auto_cast & tracer

* support master weight in dygraph for pure fp16

* check mix dtype of fp16&fp32 for check_finite_and_unscale op

* change pure fp16 funtion name

* refine some bug in auto_cast

* refine auto_cast interface logic

* add param _casted_by_pure_fp16 for class Layer

* support state_dict hook for save model by user appointed dtype in pure_fp16_decorator

* refine pure_fp16_decorator as decorator

* add unittest

* add comment

* add comment

* support recompute

* add comment for auto_cast and decorator

* support to_static_state_dict for paddle.jit.save

* unlimite models num and optimizers num

* add lookup_table in black_list

* fix momentum and layer state_dict

* fix bug in layer state_dict

* fix bug in layer state_dict_helper

* refine unittest

* refine test_momentun_op

* refine interface and some code

* refine amp_decorator interface

* refine pure fp16 interface

* refine master weight interface

* dygraph amp support param_group

* remove unused code

* fix doc

* add state_dict and load_state_dict and unittest for class GradScaler

* refine unittest for coverage of load_state_dict

* refine comments of code-block

* refine some comments

* refine state_dict code and unittest

* add #require gpu, xpu for GradScaler get/set example code

* add #require gpu, xpu for GradScaler get/set example code

* refine example code

* refine unittest for state_dict

* refine unittest for state_dict

* fix bug of DataLoader in TestGradScalerStateDict

* add flag FLAGS_cudnn_deterministic

* Support Mixed Precision training in @to_static

* fix block.vars logic

* fix GPU training loss diff

* remove unused code

* cache core.ops, test=develop

* refine, test=develop

* reduce sum op default fp32, add into amp black list

* reduce_sum default fp32 can avoid return inf when the sum value large than 65504

* add get and set for Grad_scaler

* refine some API name and comments

* refine API name and comments

* refine some comments

* Combine amp and qat
* add unit test

* add matmtl_v2 to amp list

* support dygraph

Update amp_check_finite_and_scale_op and add an updating_loss_scaling op for static graph amp training. (#26240)

* update amp_check_finite_and_scale_op for static_amp.

* use amp_check_finite_and_scale in static graph amp.

* update grads to zero when grads own infinite values(as for amp_checkout_finite_and_scale op).

* add update_loss_scaling op in cpp.

* add update_loss_scaling_op unit test.

* update the doc of the check_finite_and_unscale op

* Update the process of gradients updating skipping if the gradients have infinite values.

* update the way to zero grads.

* update test_update_loss_scaling_op.py

* add log info when find infinite grads.

* add the unit test for UpdateLossScaling Layer.

* add auto_cast, test=develop

* add loss scaler, test=develop

* add comments, test=develop

* refine code, test=develop

* refine code, test=develop

* do not set flags automatically, test=develop

* fix custom op bug, test=develop

* add more test, test=develop

* refine enable logic, test=develop

* enable amp test with GPU, test=develop

* add unittest

* add test for found_inf

* follow comments

* follow comments

* remove global variable, use singleton

* add some notes

* update comments

* update comments

* update comments

* add use_dynamic_loss_scaling argument

* refine found_inf

* refine found_inf