remove experiment_dm.py, add insturctions for dm_control to README.md

Former-commit-id: bcf31f066104b5c2b2282c2c2800b72d1ab292fb [formerly c879b9f01e402eb60ed1ddca6cc9b4daa3648043] Former-commit-id: ab6662631145ce85aa21abf633bf2e63c53061bb
zuoxingdong · May 14, 2019 · 3ebdae9 · 3ebdae9
1 parent 3a6b4f8
commit 3ebdae9
Show file tree

Hide file tree

Showing 4 changed files with 22 additions and 291 deletions.
diff --git a/baselines/README.md b/baselines/README.md
@@ -18,3 +18,25 @@ This example includes the implementations of the following reinforcement learnin
 
 ## RL
 <img src='benchmark_rl.png' width='100%'>
+
+## FAQ:
+- How to train with [dm_control](https://github.com/deepmind/dm_control) environments?
+    - Modify `experiment.py`: use [dm2gym](https://github.com/zuoxingdong/dm2gym) wrapper, e.g.
+    ```python
+    from gym.wrappers import FlattenDictWrapper
+    from dm_control import suite
+    from dm2gym import DMControlEnv
+
+    config = Config(
+        ...
+        'env.id': Grid([('cheetah', 'run'), ('hopper', 'hop'), ('walker', 'run'), ('fish', 'upright')]),
+        ...
+        )
+
+    def make_env(config, seed):
+        domain_name, task_name = config['env.id']
+        env = suite.load(domain_name, task_name, environment_kwargs=dict(flat_observation=True))
+        env = DMControlEnv(env)
+        env = FlattenDictWrapper(env, ['observations'])
+        ...
+    ```
diff --git a/baselines/ddpg/experiment_dm.py b/baselines/ddpg/experiment_dm.py
diff --git a/baselines/sac/experiment_dm.py b/baselines/sac/experiment_dm.py
diff --git a/baselines/td3/experiment_dm.py b/baselines/td3/experiment_dm.py