Initial commit

README.md

@@ -16,7 +16,7 @@ model-index:
       type: SpaceInvadersNoFrameskip-v4
     metrics:
     - type: mean_reward
-      value: 550.50 +/- 297.29
+      value: 692.50 +/- 249.73
       name: mean_reward
       verified: false
 ---
@@ -63,7 +63,7 @@ python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f lo
 
 ## Hyperparameters
 ```python
-OrderedDict([('batch_size', 32),
+OrderedDict([('batch_size', 64),
              ('buffer_size', 100000),
              ('env_wrapper',
               ['stable_baselines3.common.atari_wrappers.AtariWrapper']),
@@ -71,9 +71,9 @@ OrderedDict([('batch_size', 32),
              ('exploration_fraction', 0.1),
              ('frame_stack', 4),
              ('gradient_steps', 1),
-             ('learning_rate', 0.0001),
-             ('learning_starts', 100000),
-             ('n_timesteps', 1000000.0),
+             ('learning_rate', 1e-05),
+             ('learning_starts', 10000),
+             ('n_timesteps', 300000.0),
              ('optimize_memory_usage', False),
              ('policy', 'CnnPolicy'),
              ('target_update_interval', 1000),

args.yml

@@ -2,7 +2,7 @@
 - - - algo
     - dqn
   - - conf_file
-    - dqn.yml
+    - dqn2.yml
   - - device
     - auto
   - - env
@@ -20,7 +20,7 @@
   - - hyperparams
     - null
   - - log_folder
-    - logs
+    - logs/
   - - log_interval
     - -1
   - - max_total_trials
@@ -56,7 +56,7 @@
   - - save_replay_buffer
     - false
   - - seed
-    - 745434821
+    - 261001217
   - - storage
     - null
   - - study_name
@@ -66,7 +66,7 @@
   - - track
     - false
   - - trained_agent
-    - ''
+    - /content/logs/dqn/SpaceInvadersNoFrameskip-v4_1/best_model.zip
   - - trial_id
     - null
   - - truncate_last_trajectory

config.yml

@@ -1,6 +1,6 @@
 !!python/object/apply:collections.OrderedDict
 - - - batch_size
-    - 32
+    - 64
   - - buffer_size
     - 100000
   - - env_wrapper
@@ -14,11 +14,11 @@
   - - gradient_steps
     - 1
   - - learning_rate
-    - 0.0001
+    - 1.0e-05
   - - learning_starts
-    - 100000
+    - 10000
   - - n_timesteps
-    - 1000000.0
+    - 300000.0
   - - optimize_memory_usage
     - false
   - - policy

dqn-SpaceInvadersNoFrameskip-v4.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9caaabd9b04d36756dbfb091f04c07e2fdb8f9e75beb599813d27def2a528b9b
-size 27219597
+oid sha256:2a34eb80aa61b9175a9f08ee84bb206e6ef7ffe3ebb6e577c51af6374f67924a
+size 27219591

dqn-SpaceInvadersNoFrameskip-v4/policy.optimizer.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:95dc7cb6e07c30e5538c7bf0ddefaf8f7588f3ca92a4c26ad4943b7fc68b56b7
+oid sha256:c212a7310277932ebd1158b47086c5e2ec62e1232d479f041ee49f84d2df7103
 size 13506569

dqn-SpaceInvadersNoFrameskip-v4/policy.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c9247151d0d082f77da9d9d11d34f7de104d3f31774caf0784ea01b255abca2
+oid sha256:8be53bbeff2c10358c0a4ee5998d0cb445c5414e7a7652c02f5132b749698381
 size 13505767

results.json

@@ -1 +1 @@
-{"mean_reward": 550.5, "std_reward": 297.2915235925841, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2025-12-01T00:47:15.476379"}
+{"mean_reward": 692.5, "std_reward": 249.7323567341645, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2025-12-01T01:49:37.517283"}

train_eval_metrics.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:746a89dd276548cdf82915ada7e08a79a81faf90a110e29d8c8219c1b262f1cb
-size 36914
+oid sha256:31c6af52035b34bbe72e839be6368621ab9a00fec547514cf21f273b0eb5d950
+size 9296