…_syncbn (#27108)

* fix 2.0api, test=develop

* fix, test=develop

* Restore file changes caused by pre-commit and test it, test=document_fix

* Change argument of checkout, test=document_fix

* Remove test code, test=document_fix

* fix typo for interp_v2,test=develop

* align with torch, test=develop

* add area mode, test=develop

* fix bug, test=develop

* format notes, test=develop

* update for converage, test=develop

* fix bilinear, test=develop

* fix bicubic, test=develop

* fix typo, test=develop

* fix coverage, test=develop

* fix helper.input_dtype, test=develop

* polish notes, test=develop

* polish notes, test=develop

* polish notes, test=develop

* fix unsqueeze in dygraph

* add test

* add test

fix the error message for the unpooling.cc

* delate some wrong message test=develop, test=document_fix

* delate some wrong message

* fix weight

* fix weight and fix doc

* fix embeeding padding idx

* add UT

* fix interval

…ta cache, test=develop (#27142)

* add correlation api to contrib

* test=develop, add ps auto

* add timeout unittests retry

* modifed parameter use

…5316)

* enhance collect_op for dygraph, test=develop

* enhance detection ops with lod, test=develop

* support none bbox left in generate_proposals, test=develop

* unfiy MultiLevelRoisNum, test=develop

* update core.ops, test=develop

* add op register for new input & output, test=develop

* polish one line error message in tensor.cc

* polish error messages in tensor.cc,tensor.h tensor_impl.h

* polish error messages in tensor.cc tensor.h tensor_impl.h

* polish error messages in tensor.cc,tensor.h tensor_impl.h

* polish error messages in tensor.cc tensor.h tensor_impl.h tensor_test.cc

* polish error messages in tensor.cc tensor.h tensor_impl.h

* fix summary rnn

* fix kl_div bug when input shape is [1] and reduction is batchmean

* fix frobenius_norm error, rm p=0 2-axis support. test=develop

…led (#27035)

* resolve the issue of curl having same exit code with paddle build failed

* modifed the value of exit code

* simplify the code

Co-authored-by: chalsliu <45041955+chalsliu@users.noreply.github.com>

optimize the error message for math dir

* Add naming rule if not specific InputSpec.name

* fix function name typo

* refine comment

* remove print statement

* update requirements.txt

* add unittest_py directory and add new requirements.txt file,test=document_fix

* add unittest_py directory and add new requirements.txt file,test=document_fix

* set default value to strategy in distributed_optimizer test=develop

…7093)

* refine launch and distributed repr string for print

…is different in PY3.8 (#27201)

add lars to fleet meta optimizer

update error info for selected_rows_functor

Decrease the number of running iterations to reduce CI time.

CI system shows it decreased the unittest time from about 90 seconds to about 30 seconds

…7185)

 fix the CudaPinMemory bug for the equal op and add the test case for the equal op

* parameter_server_optimizer support auto_strategy

* add some unittest cases ot verify jit.save, no_test

* add more unittests

* add test with example inputs

* polish implement details

* remove useless blank

* fix fetch random error

* - introducing oneDNN 1.6

test=develop

* - Removed redundant code

test=develop

… test_api_impl, test=develop (#27218)

* add double grad for expand, test=develop

* add double grad for tile, test=develop

* add double grad for expand_v2 op, test=develop

…_bn_act_pass. (#27234)

* Use the single GPU card to execute the test_fuse_bn_act_pass UT.

* support to_static(model)

* add warning and unittest

* fix calcu_gradients

* fix code place

* fix embedding interface usage

* temporally disable zero_copy

* add test

* follow comments

* Add details to the summary and test it, test=document_fix

* Add set +e before example, test=document_fix

* Remove test code, test=document_fix

* Optimize summary information and test it, test=document_fix

* Remove test code, test=document_fix

* Fix the lookup_table_v2 failed on GRU mkldnn kernel issue
test=develop

* fix according to reviews, removed x_num_col_dims
test=develop

* update gru model. change according to reviews
test=develop

* change according to reviews
test=develop

Error description optimize for math dir

* Add int8 GRU kernel with UTs

* Lint fixes

* More lint fixes

* rm auto from localsgd

… 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.

* move worker loop to top level

* move reader process loop to top level

* fix failed unittests

* add check for sparse parameters with weight_decay

* move sparse check to adam.py

Enhance the error messages for  files in operators/math

Fix conv deepwise bug when in_channels=1.

* bug fix, test=develop

* improve err report, test=develop

* fix trt_dynamic_shape_ernie_deserialize_test

* support when opt cache dir does not exist

* Remove the cache in post_traning_quantization, test=develop

* replace sequence length attr to input

* optimize errror report

* add test case for pad op converter

* fix some spelling mistake commented by peiyang

* polish framework error msg part 6

* polish lossed item

* fix failed unittest

* polish by reviewer comments

…lop (#27300)

Add *.bat file for building compile environment on windows

* optimize slice TRT plugin

This patch removes unnecessary barrier for data transfer of needed offset,
so data transfer can be overlap with GPU kernel execution.

This patch also fixes incorrect name of slice plugin. That is, replaces
"layernorm" with "slice"

test=develop

* add serialize/deserialize to slice plugin

* add static shape slice trt plugin

* fix slice trt op convertor dynamic shape bug

* fix format by clang-format

* fix pylint format error

* fix problems commented by peiyang

Co-authored-by: Ryan Jeng <rjeng@nvidia.com>

* fix strategy, test=develop

* fix can_apply