Initial commit

README.md

@@ -16,7 +16,7 @@ model-index:
       type: SpaceInvadersNoFrameskip-v4
     metrics:
     - type: mean_reward
-      value: 217.50 +/- 61.98
+      value: 390.50 +/- 145.92
       name: mean_reward
       verified: false
 ---
@@ -62,23 +62,22 @@ python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f lo
 
 ## Hyperparameters
 ```python
-OrderedDict([('batch_size', 32),
-             ('buffer_size', 10000),
+OrderedDict([('batch_size', 64),
+             ('buffer_size', 1000000),
              ('env_wrapper',
               ['stable_baselines3.common.atari_wrappers.AtariWrapper']),
-             ('exploration_final_eps', 0.05),
-             ('exploration_fraction', 0.1),
+             ('exploration_final_eps', 0.01),
+             ('exploration_fraction', 0.025),
              ('frame_stack', 4),
-             ('gamma', 0.8),
+             ('gamma', 0.99),
              ('gradient_steps', 1),
-             ('learning_rate', 0.0001),
-             ('learning_starts', 10000),
+             ('learning_rate', 5e-05),
+             ('learning_starts', 50000),
              ('n_timesteps', 100000.0),
-             ('optimize_memory_usage', False),
+             ('normalize', False),
              ('policy', 'CnnPolicy'),
-             ('target_update_interval', 1000),
-             ('train_freq', 4),
-             ('normalize', False)])
+             ('target_update_interval', 10000),
+             ('train_freq', 4)])
 ```
 
 # Environment Arguments

args.yml

@@ -54,7 +54,7 @@
   - - save_replay_buffer
     - false
   - - seed
-    - 1458835778
+    - 2957842650
   - - storage
     - null
   - - study_name

config.yml

@@ -1,31 +1,31 @@
 !!python/object/apply:collections.OrderedDict
 - - - batch_size
-    - 32
+    - 64
   - - buffer_size
-    - 10000
+    - 1000000
   - - env_wrapper
     - - stable_baselines3.common.atari_wrappers.AtariWrapper
   - - exploration_final_eps
-    - 0.05
+    - 0.01
   - - exploration_fraction
-    - 0.1
+    - 0.025
   - - frame_stack
     - 4
   - - gamma
-    - 0.8
+    - 0.99
   - - gradient_steps
     - 1
   - - learning_rate
-    - 0.0001
+    - 5.0e-05
   - - learning_starts
-    - 10000
+    - 50000
   - - n_timesteps
     - 100000.0
-  - - optimize_memory_usage
+  - - normalize
     - false
   - - policy
     - CnnPolicy
   - - target_update_interval
-    - 1000
+    - 10000
   - - train_freq
     - 4

dqn-SpaceInvadersNoFrameskip-v4.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6dd869026b133ba1328f43820b459e6ff24b85a2a878ddb1f6d23dd5642985bf
-size 27220125
+oid sha256:38d005501e227a0f29d2655e0fbc99ed1bb0667086241662842a33f51074c39c
+size 27220129

dqn-SpaceInvadersNoFrameskip-v4/policy.optimizer.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:56b15d0749d2cc7f9b993b38022fadd10069037b5210e64f83c893678eb70b3a
+oid sha256:0db2e3b8c2168aa9ea8430b472d56db222b82461c9a8182eead1c057309a19bb
 size 13506236

dqn-SpaceInvadersNoFrameskip-v4/policy.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c496751a1fd0205fb387c07ce76116add5145cc906136b07f856473beb6af800
+oid sha256:14cb70e22cd28ab4c9c6d60aa9439f46666e95a3479c04910dd3d3933a753560
 size 13505370

replay.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c82371d6bc7d1b89389e744d5f8fb8669450fe263a6fe4132c1ab3cb9664d8c9
-size 264302
+oid sha256:668f5267ab783777567b8c7f33c27b372a8b413e0a52bd31f23ecc08b00dba5e
+size 257293

results.json

@@ -1 +1 @@
-{"mean_reward": 217.5, "std_reward": 61.97781861279082, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2024-01-01T02:13:36.609591"}
+{"mean_reward": 390.5, "std_reward": 145.91864171516949, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2024-01-01T16:41:16.100047"}

train_eval_metrics.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:41fefb8aa27a36c423cdccfe4c42383c31be9fa93caabf02a80e83e27c43605d
-size 5275
+oid sha256:22f2ae568f607e7d6d4aea704e318b1f5c7221fc4b47bec5c9e0e5208fe4249f
+size 5374