Add stable-baselines SAC version

araffin · Dec 28, 2018 · 7fbe91a · 7fbe91a
1 parent 47c6c1c
commit 7fbe91a
Show file tree

Hide file tree

Showing 7 changed files with 64 additions and 505 deletions.
diff --git a/docker/env_cpu.yml b/docker/env_cpu.yml
@@ -121,7 +121,7 @@ dependencies:
     - seaborn==0.8.1
     - scikit-learn==0.19.1
     - scipy==1.0.0
-    - stable-baselines==2.3.0
+    - -e git+https://github.com/hill-a/stable-baselines.git@f2890e0b02929de05e33d185f71602bf9ec4fcf9#egg=stable_baselines
     - termcolor==1.1.0
     - torchfile==0.1.0
     - tornado==4.5.3

diff --git a/docs/changelog.rst b/docs/changelog.rst
@@ -5,11 +5,13 @@ Changelog
 
 For download links, please look at `Github release page <https://github.com/araffin/robotics-rl-srl/releases>`_.
 
-Pre-Release 1.0.1 (WIP)
+Pre-Release 1.1.0 (WIP)
 -----------------------
 
 - fixed a bug in the dataset generator where the GUI was instantiated two times
 - updated stable-baselines version + srl-zoo submodule
+- add stable-baselines SAC version
+- remove pytorch SAC version **breaking changes**
 
 Release 1.0 (2018-10-09)
 -------------------------

diff --git a/environment.yml b/environment.yml
@@ -120,7 +120,7 @@ dependencies:
     - seaborn==0.8.1
     - scikit-learn==0.19.1
     - scipy==1.0.0
-    - stable-baselines==2.3.0
+    - -e git+https://github.com/hill-a/stable-baselines.git@f2890e0b02929de05e33d185f71602bf9ec4fcf9#egg=stable_baselines
     - termcolor==1.1.0
     - torchfile==0.1.0
     - tornado==4.5.3

diff --git a/rl_baselines/models/sac_models.py b/rl_baselines/models/sac_models.py
diff --git a/rl_baselines/registry.py b/rl_baselines/registry.py
@@ -25,7 +25,7 @@
     "ppo1":         (PPO1Model, AlgoType.REINFORCEMENT_LEARNING, [ActionType.DISCRETE, ActionType.CONTINUOUS]),
     "ppo2":         (PPO2Model, AlgoType.REINFORCEMENT_LEARNING, [ActionType.DISCRETE, ActionType.CONTINUOUS]),
     "random_agent": (RandomAgentModel, AlgoType.OTHER, [ActionType.DISCRETE, ActionType.CONTINUOUS]),
-    "sac":          (SACModel, AlgoType.REINFORCEMENT_LEARNING, [ActionType.DISCRETE, ActionType.CONTINUOUS]),
+    "sac":          (SACModel, AlgoType.REINFORCEMENT_LEARNING, [ActionType.CONTINUOUS]),
     "trpo":         (TRPOModel, AlgoType.REINFORCEMENT_LEARNING, [ActionType.DISCRETE, ActionType.CONTINUOUS])
 }