Initial commit

README.md

@@ -16,7 +16,7 @@ model-index:
       type: SpaceInvadersNoFrameskip-v4
     metrics:
     - type: mean_reward
-      value: 274.50 +/- 31.50
+      value: 261.50 +/- 41.42
       name: mean_reward
       verified: false
 ---
@@ -62,20 +62,20 @@ python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f lo
 
 ## Hyperparameters
 ```python
-OrderedDict([('batch_size', 64),
-             ('buffer_size', 100000),
+OrderedDict([('batch_size', 16),
+             ('buffer_size', 10000),
              ('env_wrapper',
               ['stable_baselines3.common.atari_wrappers.AtariWrapper']),
-             ('exploration_final_eps', 0.01),
-             ('exploration_fraction', 0.25),
+             ('exploration_final_eps', 0.02),
+             ('exploration_fraction', 0.1),
              ('frame_stack', 4),
-             ('gradient_steps', 2),
-             ('learning_rate', 0.0001),
+             ('gradient_steps', 1),
+             ('learning_rate', 0.001),
              ('learning_starts', 100000),
-             ('n_timesteps', 100000.0),
+             ('n_timesteps', 1000000.0),
              ('optimize_memory_usage', False),
              ('policy', 'CnnPolicy'),
              ('target_update_interval', 1000),
-             ('train_freq', 8),
+             ('train_freq', 4),
              ('normalize', False)])
 ```

args.yml

@@ -54,7 +54,7 @@
   - - save_replay_buffer
     - false
   - - seed
-    - 1380014744
+    - 3090249662
   - - storage
     - null
   - - study_name

config.yml

@@ -1,24 +1,24 @@
 !!python/object/apply:collections.OrderedDict
 - - - batch_size
-    - 64
+    - 16
   - - buffer_size
-    - 100000
+    - 10000
   - - env_wrapper
     - - stable_baselines3.common.atari_wrappers.AtariWrapper
   - - exploration_final_eps
-    - 0.01
+    - 0.02
   - - exploration_fraction
-    - 0.25
+    - 0.1
   - - frame_stack
     - 4
   - - gradient_steps
-    - 2
+    - 1
   - - learning_rate
-    - 0.0001
+    - 0.001
   - - learning_starts
     - 100000
   - - n_timesteps
-    - 100000.0
+    - 1000000.0
   - - optimize_memory_usage
     - false
   - - policy
@@ -26,4 +26,4 @@
   - - target_update_interval
     - 1000
   - - train_freq
-    - 8
+    - 4

dqn-SpaceInvadersNoFrameskip-v4.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ca880fa3efba07161dc058a306329e0ac3e4b89fb728c0bef4b50da026094914
-size 13719757
+oid sha256:d885e48cdc50aca7c8b020d05c8e5bb5c858b51dd0f77e006a6d8eb571cea2be
+size 27224402

dqn-SpaceInvadersNoFrameskip-v4/policy.optimizer.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f1e067afe9912f3dd1b7925918b8cbe439229f6008e572c9c7e431ae731419f1
-size 687
+oid sha256:12d532259f6f95e095f87f47597cc9c81ec8447d5cb62f9b812476da748ecf8c
+size 13505419

dqn-SpaceInvadersNoFrameskip-v4/policy.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:24514105126a149636622f02550979a5ea7d74c27de67f1760dd68ee6c30b6c9
-size 13504937
+oid sha256:c407119b70d4f9af5a8b3c95880361b767fe6d393f05982607a821c357ba9e26
+size 13504745

dqn-SpaceInvadersNoFrameskip-v4/system_info.txt

@@ -2,6 +2,6 @@
 - Python: 3.9.16
 - Stable-Baselines3: 1.8.0a9
 - PyTorch: 1.13.1+cu116
-- GPU Enabled: True
-- Numpy: 1.22.4
+- GPU Enabled: False
+- Numpy: 1.24.2
 - Gym: 0.21.0

replay.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fba2ab7031f7061f065c3d5bc1de3dcfb9536d40b9bccd2413c48c7909f59ea3
-size 261744
+oid sha256:e7267eeb72f97d23e26e79a38f73024988ea5fc5381f7fcd55c0ef651ef4e38d
+size 263544

results.json

@@ -1 +1 @@
-{"mean_reward": 274.5, "std_reward": 31.5, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2023-03-20T11:06:20.310249"}
+{"mean_reward": 261.5, "std_reward": 41.415576779757636, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2023-03-20T18:58:35.003220"}

train_eval_metrics.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:211b9f56f9962bf7e6f95cef3b8bf226e37ef74dcf6d0cbfbb4fe7b478b7df62
-size 5803
+oid sha256:06cd0586a1c94455c3309573b64159908ea1052d32e60858b86602fdc6fbafd9
+size 41552