add envs

examples/OpenAIGym/README.md

 
-# To run pretrained atari model for 100 episodes:
+# To run a pretrained Batch-A3C atari model for 100 episodes:
 
 1. install [tensorpack](https://github.com/ppwwyyxx/tensorpack)
 2. Download models from [model zoo](https://drive.google.com/open?id=0B9IPQTvr2BBkS0VhX0xmS1c5aFk)
@@ -15,7 +15,16 @@ Models are available for the following gym atari environments (click links for v
 + [Asteroids-v0](https://gym.openai.com/evaluations/eval_8eHKsRL4RzuZEq9AOLZA)
 + [Atlantis-v0](https://gym.openai.com/evaluations/eval_Z1B3d7A1QCaQk1HpO1Rg)
 + [Breakout-v0](https://gym.openai.com/evaluations/eval_L55gczPrQJamMGihq9tzA)
++ [DemonAttack-v0](https://gym.openai.com/evaluations/eval_tt21vVaRCKYzWFcg1Kw)
++ [DoubleDunk-v0](https://gym.openai.com/evaluations/eval_FI1GpF4TlCuf29KccTpQ)
++ [ElevatorAction-v0](https://gym.openai.com/evaluations/eval_SqeAouMvR0icRivx2xprZg)
++ [FishingDerby-v0](https://gym.openai.com/evaluations/eval_pPLCnFXsTVaayrIboDOs0g)
 + [Pong-v0](https://gym.openai.com/evaluations/eval_8L7SV59nSW6GGbbP3N4G6w)
 + [Seaquest-v0](https://gym.openai.com/evaluations/eval_N2624y3NSJWrOgoMSpOi4w)
++ [Tennis-v0](https://gym.openai.com/evaluations/eval_gDjJD0MMS1yLm1T0hdqI4g)
++ [VideoPinball-v0](https://gym.openai.com/evaluations/eval_PWwzNhVFR2CxjYvEsPfT1g)
 
-Note that atari game settings in gym is quite different from DeepMind papers, so the scores are not comparable.
+Note that atari game settings in gym are quite different from DeepMind papers, so the scores are not comparable. The most notable differences are:
++ In gym, each action is randomly repeated 2~4 times.
++ In gym, inputs are RGB instead of greyscale.
++ In gym, an episode is limited to 10000 steps.

examples/OpenAIGym/run-atari.py

@@ -76,7 +76,7 @@ def run_submission(cfg):
     dirname = 'gym-submit'
     player = get_player(dumpdir=dirname)
     predfunc = get_predict_func(cfg)
-    for k in range(10):
+    for k in range(100):
         if k != 0:
             player.restart_episode()
         score = play_one_episode(player, predfunc)