* Add fusion_lstm INT8 PTQ

* Correct mkldnn_cache_capacity and enable fc_lstm_fuse_pass only for this test

* Change mkldnn_cache_capacity

* add pool2d grad grad

* dbg

* add unittest

* update format

* add more unittests

* dbg

* transpose/slice/stride_slice/squeeze/unsqueeze op_def_enhance-1.0

* delete infer_flags and decrease-axis

* delet infer_flags and decrea_axis

* add pool2d pool3d extra()

* delete ceil_mode extra()

* delete ceil_mode extra()

* delete ceil_mode extra()

* add extra to use_mkldnn

* add kernel, stride check

* add unitest for param out of range

* delete max limit check

* Support Reset for DeviceEvent

* fix code

* add more unittest

* add AsExtra tag for conv transpose op

* check the existence of use_cudnn before get this attribute

* add depthwise_conv_npu_grad op

* add depthwise_conv_npu_grad op

* add depthwise_conv_npu_grad op

* add NHWC test case

* support numpy dtype and polish code of list index.

* polish code.

This PR adds error exception in fused transformer python interface.
The function body are not implemented (will be implemented later).
Following zhiqiu's comment in previous PR-35206 (merged already), it is better to raise an exception instead of using "pass".

* where op add broadcast, when x_shape == y_shape, and x_shape != cond_shape

* add static api tests, and delete debug codes

* modify gc logic, use new device_event, test=develop

* use GenerateDeviceEventFlag, test=develop

* refine, test=develop

* fix test_standalone_executor.py, test=develop

* refine, test=develop

* support lodtensorarray

* [NPU] Add huber_loss op

* [NPU] Add huber_loss op

* [NPU] Add huber_loss p[

* [NPU] Add huber_loss

* [NPU] add int64 kernels, test=develop

* update ci scripts to be able to trun WITH_ASCEND_INT64 on, test=develop

* Add AsExtra for transpose

* add AsExtra for lstm op

* add AsExtra for gru

* [iscan] bugfix

* test_standalone_executor modify

* add conv2d_transpose npu op

* CopyRight 2020 to 2021

* add fp32

* delete repeat test case

* delete repeat test case

* fix paddle.NPUPlace

add AsExtra to the following attribute of partial_sum op:
1. use_mkldnn

add AsExtra to the  following attribute of concat op:
  1. use_mkldnn
  2. use_quantizer
  3. mkldnn_data_type

add AsExtra to the use_mkldnn attribute of reshape op

* [NPU] update npu ut skil scirpts, test=develop

* update, test=develop

* update scripst, test=develop

* add debug code, test=develop

* update

* add lookup_table_v2_op AsExtra

* add grap attr asextra

* add gather_op asExtra()

* add lookup_table_v2_op AsExtra

* add grap attr asextra

* add elementwise_pow_grad_npu

* fixed bug for CI

* deal with comments

* fixed bug for CI

* deal with comments

* add log_softmax_op_npu

* log_softmax_op_v1

* import test_log_softmax_grad

* [NPU] Support npu kernel for gather_ng op

* [NPU] Support npu kernel for gather_nd op

* [NPU] Support npu kernel for gather_nd and gather_nd_grad op

* update py format error.

* modify gather_nd_op_npu

* modify gather_nd 910 test

* modify gather_nd 910 test
Co-authored-by: Nxiaoxiaohehe001 <hiteezsf@163.com>