- removed vf clipping in pposgd - that was severely degrading performance on mujoco because it didn’t account for scale of returns
- switched adam epsilon in pposgd_simple
- brought back no-ops in atari wrapper (oops)
- added readmes
- revamped run_X_benchmark scripts to have standard form
- cleaned up DDPG a little, removed deprecated SimpleMonitor and non-idiomatic usage of logger

* Export param noise

* Update documentation

* Final finishing touches

ppo and trpo

Effectively apply weights from the replay buffer

Fix TF graph variables deprecation

It seems that the weights retrieved from the replay buffer are not applied when training the model. Is there any reason for that or am I missing something?

In any case, I have added a parameter in order for them to be used; just in case it is useful.

remove unnecessary initialization of variable resized_screen

Update setup.py

Fix README since BreakOut pretrained model doesn't match the correct …

Fix README since BreakOut pretrained model doesn't match the correct tensor shape. Therefore, Pong is used instead.

docstring and comment fixes

Update README.md

Consistent initial type (float) for episode_rewards

Update setup.py

`PongNoFrameskip-v4` seems to require `gym>=0.9.1`

Fix typos

tf_util: spelling / docstring updates

Correct path to script "download_model"

`python -m baselines.deepq.experiments.download_model` becomes `python -m baselines.deepq.experiments.atari.download_model`