* Test examples for pep8 compliance.
* Make rl_pong example pep8 compliant.
* Make policy gradient example pep8 compliant.
* Make lbfgs example pep8 compliant.
* Make hyperopt example pep8 compliant.
* Make a3c example pep8 compliant.
* Make evolution strategies example pep8 compliant.
* Make resnet example pep8 compliant.
* Fix.
* pong on pixels working (not cleaned up)
* make training compatible with all atari games
* cartpole runs
* Update documentation and usage for policy gradients.