change log dir

tf_agents_test.py

@@ -18,7 +18,7 @@ r"""Train and Eval DDPG.
 To run:
 
 ```bash
-tensorboard --logdir $HOME/tmp/ddpg/gym/HalfCheetah-v2/ --port 2223 &
+tensorboard --logdir ./data/tmp/ddpg/gym/HalfCheetah-v2/ --port 2223 &
 
 python tf_agents/agents/ddpg/examples/v2/train_eval.py \
   --root_dir=$HOME/tmp/ddpg/gym/HalfCheetah-v2/ \
@@ -66,7 +66,7 @@ from tf_agents.environments import wrappers
 from tf_agents.environments import suite_gym
 from tf_agents.trajectories import time_step as ts
 
-flags.DEFINE_string('root_dir', '$HOME/tmp/ddpg/gym/HalfCheetah-v2/',
+flags.DEFINE_string('root_dir', './data/tmp/ddpg/gym/HalfCheetah-v2/',
                     'Root directory for writing logs/summaries/checkpoints.')
 flags.DEFINE_integer('num_iterations', 100000,
                      'Total number train/eval iterations to perform.')
@@ -106,7 +106,7 @@ class CardGameEnv(py_environment.PyEnvironment):
       return self.reset()
 
     # Make sure episodes don't go on forever.
-    if action >= 5.0:
+    if action >= 1.0:
       self._episode_ended = True
     elif action < 1.0:
       new_card = np.random.randint(1, 11)