Initial Commit

README.md

@@ -32,6 +32,8 @@ with hyperparameter optimization and pre-trained agents included.
 ## Usage (with SB3 RL Zoo)
 
 RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo
+SB3: https://github.com/DLR-RM/stable-baselines3
+SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib
 
 ```
 # Download model and save it into the logs/ folder
@@ -45,3 +47,21 @@ python train.py --algo dqn --env MountainCar-v0 -f logs/
 # Upload the model and generate video (when possible)
 python -m utils.push_to_hub --algo dqn --env MountainCar-v0 -f logs/ -orga sb3
 ```
+
+## Hyperparameters
+```python
+OrderedDict([('batch_size', 128),
+             ('buffer_size', 10000),
+             ('exploration_final_eps', 0.07),
+             ('exploration_fraction', 0.2),
+             ('gamma', 0.98),
+             ('gradient_steps', 8),
+             ('learning_rate', 0.004),
+             ('learning_starts', 1000),
+             ('n_timesteps', 120000.0),
+             ('policy', 'MlpPolicy'),
+             ('policy_kwargs', 'dict(net_arch=[256, 256])'),
+             ('target_update_interval', 600),
+             ('train_freq', 16),
+             ('normalize', False)])
+```

dqn-MountainCar-v0.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c5d02d795e9120efb19cd1f65541a7f040eacf9b4253bfd438ebc35b64113a14
+oid sha256:cc7f7861b17461282de2b85694d9ca48e97a97c650513acf42021f70182f56db
 size 1103767

dqn-MountainCar-v0/data

@@ -4,15 +4,15 @@
         ":serialized:": "gAWVMAAAAAAAAACMHnN0YWJsZV9iYXNlbGluZXMzLmRxbi5wb2xpY2llc5SMCURRTlBvbGljeZSTlC4=",
         "__module__": "stable_baselines3.dqn.policies",
         "__doc__": "\n    Policy class with Q-Value Net and target net for DQN\n\n    :param observation_space: Observation space\n    :param action_space: Action space\n    :param lr_schedule: Learning rate schedule (could be constant)\n    :param net_arch: The specification of the policy and value networks.\n    :param activation_fn: Activation function\n    :param features_extractor_class: Features extractor to use.\n    :param features_extractor_kwargs: Keyword arguments\n        to pass to the features extractor.\n    :param normalize_images: Whether to normalize images or not,\n         dividing by 255.0 (True by default)\n    :param optimizer_class: The optimizer to use,\n        ``th.optim.Adam`` by default\n    :param optimizer_kwargs: Additional keyword arguments,\n        excluding the learning rate, to pass to the optimizer\n    ",
-        "__init__": "<function DQNPolicy.__init__ at 0x7fd067581b00>",
-        "_build": "<function DQNPolicy._build at 0x7fd067581b90>",
-        "make_q_net": "<function DQNPolicy.make_q_net at 0x7fd067581c20>",
-        "forward": "<function DQNPolicy.forward at 0x7fd067581cb0>",
-        "_predict": "<function DQNPolicy._predict at 0x7fd067581d40>",
-        "_get_constructor_parameters": "<function DQNPolicy._get_constructor_parameters at 0x7fd067581dd0>",
-        "set_training_mode": "<function DQNPolicy.set_training_mode at 0x7fd067581e60>",
+        "__init__": "<function DQNPolicy.__init__ at 0x7f8b5764cb00>",
+        "_build": "<function DQNPolicy._build at 0x7f8b5764cb90>",
+        "make_q_net": "<function DQNPolicy.make_q_net at 0x7f8b5764cc20>",
+        "forward": "<function DQNPolicy.forward at 0x7f8b5764ccb0>",
+        "_predict": "<function DQNPolicy._predict at 0x7f8b5764cd40>",
+        "_get_constructor_parameters": "<function DQNPolicy._get_constructor_parameters at 0x7f8b5764cdd0>",
+        "set_training_mode": "<function DQNPolicy.set_training_mode at 0x7f8b5764ce60>",
         "__abstractmethods__": "frozenset()",
-        "_abc_impl": "<_abc_data object at 0x7fd0675794e0>"
+        "_abc_impl": "<_abc_data object at 0x7f8b576444e0>"
     },
     "verbose": 1,
     "policy_kwargs": {
@@ -89,12 +89,12 @@
         ":serialized:": "gAWVNQAAAAAAAACMIHN0YWJsZV9iYXNlbGluZXMzLmNvbW1vbi5idWZmZXJzlIwMUmVwbGF5QnVmZmVylJOULg==",
         "__module__": "stable_baselines3.common.buffers",
         "__doc__": "\n    Replay buffer used in off-policy algorithms like SAC/TD3.\n\n    :param buffer_size: Max number of element in the buffer\n    :param observation_space: Observation space\n    :param action_space: Action space\n    :param device:\n    :param n_envs: Number of parallel environments\n    :param optimize_memory_usage: Enable a memory efficient variant\n        of the replay buffer which reduces by almost a factor two the memory used,\n        at a cost of more complexity.\n        See https://github.com/DLR-RM/stable-baselines3/issues/37#issuecomment-637501195\n        and https://github.com/DLR-RM/stable-baselines3/pull/28#issuecomment-637559274\n    :param handle_timeout_termination: Handle timeout termination (due to timelimit)\n        separately and treat the task as infinite horizon task.\n        https://github.com/DLR-RM/stable-baselines3/issues/284\n    ",
-        "__init__": "<function ReplayBuffer.__init__ at 0x7fd0679f2ef0>",
-        "add": "<function ReplayBuffer.add at 0x7fd0679f2f80>",
-        "sample": "<function ReplayBuffer.sample at 0x7fd0679e7680>",
-        "_get_samples": "<function ReplayBuffer._get_samples at 0x7fd0679e7710>",
+        "__init__": "<function ReplayBuffer.__init__ at 0x7f8b57abeef0>",
+        "add": "<function ReplayBuffer.add at 0x7f8b57abef80>",
+        "sample": "<function ReplayBuffer.sample at 0x7f8b57ab3680>",
+        "_get_samples": "<function ReplayBuffer._get_samples at 0x7f8b57ab3710>",
         "__abstractmethods__": "frozenset()",
-        "_abc_impl": "<_abc_data object at 0x7fd067a52480>"
+        "_abc_impl": "<_abc_data object at 0x7f8b57b1d480>"
     },
     "replay_buffer_kwargs": {},
     "train_freq": {

results.json

@@ -1 +1 @@
-{"mean_reward": -103.4, "std_reward": 7.4859869088851605, "is_deterministic": true, "n_eval_episodes": 10, "eval_datetime": "2022-05-20T09:54:22.907164"}
+{"mean_reward": -103.4, "std_reward": 7.4859869088851605, "is_deterministic": true, "n_eval_episodes": 10, "eval_datetime": "2022-05-20T09:58:38.915474"}