* add conv3d prototxt

* Update conv3d.pbtxt

* Update conv3d.pbtxt

* 1, remove layernorm dynamic fp16; 2, let reshape out in dynamic shape

* remove useless code

add the support for the bool in compare ops

* add reduce_sum_op by add self-kernel

* set all ReduceKernel MPType for accuracy

* add float16 test script which input is integer number

* solve reduce sum float16 check_grad problem

* solve conflict and change test script for CI

* change kernel register for CI

* remove all useless template

* Add digamma_op and unittest

* add digamma_op api

* remove special DigammaCudaKernel and correct some docs

* remove unused headers

* fix api doc error

* Eliminate numerical differences of LayerNorm; fix LayerNorm Nan Bug while large data input

* fix bug while large shape of data input

* Small changes related to BF16 fusion_gru and fusion_lstm

* Correct to pass arg by value

* Add conditions to rnn op

* Correct the spelling mistake

* Improving the test with checking activation

* Trigger CI

* fix gather shape bug

* fix None

* fix topo

* support diff dataset tensor place in single process dataloader

* fix unittest failed

* add parallel_cross_entropy

* add grad for crossentropy

* fix cross entropy

* add op pbtxt of matmul/matmul_v2/scale/softmax

* add newline

* move quant attrs to extra for matmul

* add two attributes for yolo box

* Create squeeze2.pbtxt

* Create unsqueeze2.pbtxt

* Update unsqueeze2.pbtxt

* Update squeeze2.pbtxt

* add transpose transpose opdef, test=develop

* add line, test=develop

* add pool2d and reshape2 pbtxt

* add pool2d and reshape2 pbtxt, test=develop

* delete reshape2.txt, test=develop

* add lines, test=develop

* replace 'InnerSetOverridedStopGradient' with 'SetOverridedStopGradient'.

* improve coverage.

* polish error message.

* add dynamic layer_norm plugin

* fix bug

* fix numpy.allclose

* fix format

* fix code style

* remove shepe in dynamic shape

* code format

* remove layer norm fp16

* fix format

* Fix two english api documents, transpose and strided_slice

* OP:strided_slice_op supports bool type inputs

* add op_def for gru, lstm and layer_norm

* fix no newline flag

* Add the op def for elementwise_div elementwise_pow elementwise_sub fake_quantize_dequantize_moving_average_abs_max reduce_mean sqrt test=develop

* remove .ipynb_checkpoints for opdef

* remove fake_quantize_dequantize_moving_average_abs_max for opdef

* fix c_split bug

* fix utest

* add c_embedding for tensorparallel

This reverts commit 7be6191b.

* fix inference prepare data bug

* rename functions

* typo

* typo

* typo

* UT correct

* correct condition

* correct condition

* ci coverage

* morelines

* fix ci coverage

This PR made these changes to support double grad:

1. Translate `paddle.grad` to `paddle.static.gradients` to support double grad for dy2stat.
2. Fix IfElseTransformer bug which may not change value if "Store before Load" variable is in "Store" statement is in IfElse conditional statement
3. Add `DOut` to support double grad variables in `run_program_op`
4. Add support for renaming for double grads for `jit.save/load`