* shape and optimize paddle error message stack, test=develop

* limit exception type & add unittest, test=develop

* fix multi-platform problem, test=develop

* fix related unnitest failed, test=develop

* add doc & fix unittest errors, test=develop

* fix function name error, test=develop

* update tensor test exception msg compare, test=develop

* remove unittest on win32, the dir format is different, test=develop

* remove useless package, test=develop

* add paddle enforce handler unittest, test=develop

* add exception checkout, test=develop

* fix coverage failed, test=develop

* fix op registry test failed, test=develop

* refactor whole pr, test=develop

* remove test in CMakelist, test=develop

* fix coverage, test=develop

Use AllocateTmpTensor() for creating temporary tensors in warpctc.

Set output type LoDTensor only

After code experiment, I found data doens't support other type

* fix DistributeTranspilerConfig document, test=develop

* Make PaddleSlim support PyReader.
* Fix unittest of sensitive pruning.
* Add some assert.

test=develop

* Fix conv2d+dequantize squash for residual fusion

test=develop

* Correct int8 input

test=develop

* Add if exclude or include padding in pool2d mkldnn

test=develop

test=develop

* x.dims == y.dims test=develop

* refine comment

* Expose `mutable_data` as python binding

test=develop

* Add test for device pointer binding

test=develop

* Make test compatible with python 2

* fix input shape check test=develop

* move PADDLE_ENFORCE test=develop

Add dtype for coalesce_tensor_op

test=develop
test=document_fix

Polish elementwise max min pow document to add more examples

test=develop

* fix doc of apply_optimize
test=document_fix
test=document_preview

* modify doc of backward
test=develop
test=document_fix

* modify document hash
test=develop
test=document_preview

* add lod check for sequence op, test=develop

* delete unnecessary check in expend op, test=develop

The new "fluid.data" changes old "fluid.layers.data":

1. Add shape and dtype check.
2. Remove "append_batch_size" parameter. We won't offer this in the new data layer because other deep learning platforms don't have this kind of data layer pre-processing. It may confuse users.
3. Remove "stop gradient" parameter because the data layer doesn't do back-propagation

TODO：
Now data layer feeded by executor is checked, will we want to check the feed data of readers in the future?

fix memory leak in HogwildWorker,  whose ops are  explicitly deleted in destructor

* add AdadeltaOptimizer doc, test=develop

* refine doc,test=develop

* folllow lanxiang's comments, test=develop, test=document_fix

* expose set_gradient_clip, test=develop, test=document_preview, test=preview

* expose gradient clip, test=develop, test=document_fix

* refine doc, test=develop

* follow lanxiang's comments, test=develop, test=document_fix

* refine doc, test=develop, test=document_preview

* add kernel for fill_op, test=develop

* modify PADDLE_ENFORCE to PADDLE_ENFORCE_EQ, test=develop

* add op test for fill_op, test=develop

* REGISTER COP CUDA KERNEL, test=develop

* update test_fill_op.py, test=develop

* change FillConstantOpVarTypeInference to FillOpVarTypeInference, test=develop

* fix op test, test=develop

* add head file, test=develop

* add support tensor and tensorlist for strided_slice OP test=develop

* fix the commnet test=develop

* fix test=develop

* fix the bug test=develop

* delete log test=develop

* fix API.spec test=develop

* fix test=develop

* update API.spec

* fix bn

* add support of matmul with multiple head even different width and height

Original matmul with multiple head supports only the mat_a.width == mat_b.height,
in that case, mat_b will be horizontally split. In this patch, we extend the
support when mat_a.width != mat_b.height but mat_a.width/head_number == mat_b.height,
in this case, mab_b will be vertically split.

One example is A is [3, 8], B is [2, 16], head_number is 4. In this
case, A will be split as [3, 2], B will be (vertically) split as
[2, 4]. The final result will be 4 matrix of 4 matrix of [3,4], i.e. [3, 16]

test=develop

* add support of matmul with multiple head even different width and height

Original matmul with multiple head supports only the mat_a.width == mat_b.height,
in that case, mat_b will be horizontally split. In this patch, we extend the
support when mat_a.width != mat_b.height but mat_a.width/head_number == mat_b.height,
in this case, mab_b will be vertically split.

One example is A is [3, 8], B is [2, 16], head_number is 4. In this
case, A will be split as [3, 2], B will be (vertically) split as
[2, 4]. The final result will be 4 matrix of 4 matrix of [3,4], i.e. [3, 16]

test=develop

* refactor the code of matmul with multiple head even different width and height

test=develop

* refine ctc align op with padding 
* refine api sample code

FIx C++ inference BUG: When open memory optim and enable trt subgraph at the same time, there is a bug (#19969)

* fix memory optimization type
test=develop

* 1. fix BUG: open trt and memory optim will trigger bug.
2. Clean memory optim bug.
test=develop

* Add support for new QAT models

test=develop
Co-Authored-By: NMichał Gallus <michal.gallus@intel.com>
Co-Authored-By: NWojciech Uss <wojciech.uss@intel.com>

* fixed fps results

test=develop

* fix top5 accuracy drop problem

* updated for new QAT models

* skip quantizing average pooling - dirty but working

* add missing pass

* added missing conv+brelu fuse pass

* removed a call to non-existent pass

test=develop

* renamed pass

test=develop

* Adjust finding pooling scale to newest QAT models

* Remove unnecessary code from quantization_mkldnn_pass

* Copy Pooling input scale to output scale in QAT

* Refactor & remove unused code in QAT

* Incorporate fp32 FC into QAT

test=develop

* Enable graph drawing with debug flag

test=develop

* Add tests for QATv2

* Fix paths for QATv2 models

test=develop

* Add option to save transformed int8 qat model

test=develop

* Remove redundant lines from qat mkldnn pass

test=develop

* Delegate disablement of avg pooling to qat

test=develop

* fix CI bug, test=develop

* Follow Wangzhen's Review, test=develop

* Update API.spec

test=develop

* Name False in (is_unsigned, TensorScale) tuple

test=develop

* Removing last dims constraints of seq_pad and seq_unpad test=develop

* fix test_layer api code test=develop

* fix sequence_pad_op.cc conflict test=develop

* remove test_analyzer_mm_dnn test=develop

* fix vectorize bug test=develop

* fix vectorize<int> test=develop

test=develop

test=develop
test=document_fix

* Add float16 support to `sync_batch_norm_op`

test=develop

* Add test for sync_bn with FP16 input

test=develop

* Remove constraint that last dimension is forced to be 1 by add
lookup_table_v2 test=develop

* modify into PADDLE_ENFORCE_CUDA_SUCCESS test=develop

* Revert "modify into PADDLE_ENFORCE_CUDA_SUCCESS test=develop"

This reverts commit 8a960bfc61e51aa27c3c529df8fb90b93ebd19f9.

* move api into fluid.embedding test=develop

* fix example code test=develop

* move one_hot into fluid.one_hot

* modify api.spec test=develop

* fix loss shape test=develop