Add refer code to get started ~

fix pipe reader on multi passes

Add ChunkEvaluator for Multi-batches

Fix typo in cross-compiling documentation of iOS and add it to the index

add ios cross compiling doc in en

Multigpu Feature

add MKLDNNPlace

Conflicts:
	paddle/platform/place.h

Expose GradOpMaker to Python

* Make cast op support bool

Also add `elemwise_sub/mul/abs/clip` layers

* Make fuild.layers as a module

* Move layers as a module

* Split layers.py into layers module

* Fix CI

* Fix CI

update the link of docs icon

Unify `step_block` and `block` to `sub_block`

* "derived cudnnDevice context"

* "leave remove cudnn handle from CUDADeviceContext"

* "fix math function error"

Enhance Parameters.init_from_tar in V2

Enhance ReduceOp to support reducing over all elements

Refine conv