Fix conv2d bias

Add dilations for conv2d and optimize conv2d code

Enable manylinux builds

* fix gitignore

* refine cmake file

* add decayed adagrad python code

* fix typo and order

* small fix

The size of the bias parameter should be the number of filters.

Update the annotations of layers.py.

 Simpily demo, add paddle.default_main_program() and paddle.default_startup_program()

Feature/evaluator

Fix/reduce op

- Removed all main_program and startup_program in the demo.
- Using paddle.default_main_program() hides the implementation detail (e.g., using g_main_program) from the user, we can change the implementation in the future much easier.

* Adding greater than and less than equal ops to compare op
* Changing the name of the less_than_equal and greater_than_equal op
* Also changing the name of the functors

Remove some code to reduce the size of the mobile inference library.

fix pooling functor parameter order

Add dilation for exconv layer

max pool Layer with mask