* support run haokanctr model in heterps-models

* polish setup.py

* polish JVM_LIB in evn_dict

* align infer auc with DistPsArch pre-stable

* async and multi thread data feed

* rewrite dense tensor intialization

* async infer shape and reuse memory

* adjust defalut tolerance of output and grad

* fix a bug in the grad of OpTest

* fix the type of setting defalut value in optest, both forward and
backward

* add defalut

* fix test_sum_op

* fix test_sum_op test for testing add_n

* modify the add_n op_test

* add scatter composite rule.

* add public_python_api

* add python unit16 support.

* fix code style.

* add cinn to makelist

* cinn unsupport uint16, forbidden cinn when dtype==uint16.

* update assign_pos

* update attention_lstm

* update barrier

* update batch_fc

* update beam_search

* update beam_search_decode

* update bilateral_slice

* fix bug

* Handle Structure kernel for InterpreterCore::RunOperator

* fix bug

* fix rocm compile

* fix rocm compile

* Revert "fix rocm compile"

* test

* revert test and update cmake

---------
Co-authored-by: Nchenruibiao <chenruibiao@baidu.com>

* Support ignore_index for c_softmax_with_cross_entropy_op.

* Polish code. Remove useless comments and add Testcase.

* Polish code for TestCase.

* Polish code.

* Polish code style.

* Polish code.

* Change loss calculation formula and ignore_index dtype.

* Polish TestCase.

* Fix bug of c_softmax_with_cross_entropy_op_xpu_op. Attribute 'ignore_index'
dtype is int64_t.

* delete check_dygraph and use default atol,max_relative_error

* add test case for bicubic_interp

* add FP16 for multinomial

* fix input data

* update code

* fix FP16

* fix code

* remove sync_calc_stream and sync_comm_stream

* fix ci bug

* fix

* fix

* fix

* support auto generate for prelu

* op_compat 中增加输入参数

* del attrs ; add kernel data_type

* add PreluGradInferMeta

* stride slice fp16 and bf16 unitest

* fix code style

* add self.dtype

* [Test Mv] ipu_test

* [Test Mv] cmake add py_test_modules

* [Move Test] rm py_test_modules

* rm asp

* add autogen code support for sigmoid_cross_entropy_with_logits

* add inplace

* add autogen code support for merge_selected_rows

* bug fixed

* force sync batch norm grad sequential

* add fp16 bfp16 test

* polish

* polish

* polish

* polish

* add type check

* fix bug in proto

* add utest

* change op with xpu

* change range yaml

* fix bug in generate_op.py

* support python object input data broadcast for model parallel

* add unittest

* fix

* fix concat 0D tensor

* fix codestyle

* [CINN] pass global seed to CINN

* fix cu not include cinn/runtime/flags.h bug

* fix DefaultCUDAGenerator should has device id bug

[CodeStyle][C416][C417] rewrite unnecessary comprehension with function call and use generator instead of map (#52140)

* codestyle c416 c417

* fix error

* fix inc

* unify all C4 rules into one

* fix inc

---------
Co-authored-by: NSigureMo <sigure.qaq@gmail.com>

* delete old dygraph op test

* add gloo  send_recv

* fix code_stype

* fix CI bug

* fix send_recv.cc

* add send_recv without sync_op

* fix send_recv test

* fix gather.cc

* pad3d add unittests of fp16 and bf16

* pad3d add unittests of fp16 and bf16

* fix cuda place

* fix random to uniform

* fix class name

* fix fp16 max relative error to 1.5e-3

* add dytpe register for onednn

* add pad uint16 check of common.py

* remove check_eager

* test_check_grad --> test_check_grad_normal

* Clear the infrt-related code

* remove tools/infrt

* Fix flashattn build error on jetson

* Fix nvcc not found on jetson

* add group_norm composite rule

* add test for scale_grad and bias_grad

* resolve conflicts

* remove amp in composite_rule.py

* add float16 test

* deal with NHWC format

* keep the composite rule in float16 identical as original kernel

* resolve conflicts