* change int to floor, test=develop

* fix unittest, test=develop

* Indent MSELoss example docs

* Point out input tensors should be of same shape

test=develop

* Document `MSELoss` input and return parameters

test=develop,test=document_fix

* QAT: support range-based quantization and scales from attribute

* added support for channelwise

* Init complex number neural network, test=develop

* Improve doc writing, test=develop

* Fix elementwise add & sub, test=develop

* Fix elementwise mul act, test=develop

* a) add ut for complex variable; b) remove arg act in elementwise_ops. test=develop

Repair unittest: test_dyn_rnn

* add gather split squeeze stack unsqueeze api test=develop

* add gather split squeeze stack unsqueeze api test=develop

* fix bug test=develop

* fix bug test=develop

* fix bug test=develop

* fix bug test=develop

* fix bug test=develop

* test=develop

* update mklml.cmake to 2019.0.5

test=develop

* update mklml.cmake with new version

test=develop

* support the head number == 1
test=develop

* fix slice op error.
test=develop

Add tracks_own_finished to Decoder to avoid mismanagement of the finished state in dynamic_decode. (#23664)

test=develop

* upload code for tensor.rand

* fix import

* update example, change paddle.tensor.rand to paddle.rand

* change 'variable' to 'Variable',  test=develop

change 'variable' to 'Variable' in description,  test=develop

* add pre-commit check

* add pre-commit check

* pre-commit check, test=develop

* add more unittest code

* trigger ci, test=develop

* pre-commit check, test=develop

* update api and test comment, test=develop

* update api and test comment, test=develop

* add more type check, test=develop

* add detail error info for device, test=develop

* add unnittest, test=develop

* resolve conflict and pre-commit check, test=develop

* Make batch_float_feasigns & batch_uint64_feasigns as member variable

* update name generator, test=develop

* use c++ unique name generator, test=develop

* Make layers.reshape/expand/slice in dygraph support var inputs.
Make transpose support size 0.
test=develop

* Update layers.expand and layers.slice to support var inputs.
test=develop

test=develop

* update docs, test=develop

* polish eng docs, test=develop

* lod_rank_table error message enhance. test=develop

Fix the dygraph mode in matmul, add the support in Linear Op

* add the thread_local_allocator, test=develop

* refactor the thread_local_allocator, test=develop

* provides option setting strategy, test=develop

* add DataLoader, Dataset, BatchSampler