Initial commit

README.md

@@ -16,7 +16,7 @@ model-index:
       type: SpaceInvadersNoFrameskip-v4
     metrics:
     - type: mean_reward
-      value: 268.50 +/- 78.17
+      value: 319.00 +/- 105.97
       name: mean_reward
       verified: false
 ---
@@ -63,20 +63,20 @@ python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f lo
 
 ## Hyperparameters
 ```python
-OrderedDict([('batch_size', 512),
-             ('buffer_size', 100000),
+OrderedDict([('batch_size', 64),
+             ('buffer_size', 50000),
              ('env_wrapper',
               ['stable_baselines3.common.atari_wrappers.AtariWrapper']),
              ('exploration_final_eps', 0.01),
              ('exploration_fraction', 0.1),
-             ('frame_stack', 4),
+             ('frame_stack', 8),
              ('gradient_steps', 1),
-             ('learning_rate', 2.5e-05),
-             ('learning_starts', 100000),
-             ('n_timesteps', 1000000.0),
+             ('learning_rate', 0.0001),
+             ('learning_starts', 50000),
+             ('n_timesteps', 200000.0),
              ('optimize_memory_usage', False),
              ('policy', 'CnnPolicy'),
-             ('target_update_interval', 1000),
+             ('target_update_interval', 64),
              ('train_freq', 32),
              ('normalize', False)])
 ```

args.yml

@@ -56,7 +56,7 @@
   - - save_replay_buffer
     - false
   - - seed
-    - 3309286548
+    - 2478037883
   - - storage
     - null
   - - study_name

config.yml

@@ -1,8 +1,8 @@
 !!python/object/apply:collections.OrderedDict
 - - - batch_size
-    - 512
+    - 64
   - - buffer_size
-    - 100000
+    - 50000
   - - env_wrapper
     - - stable_baselines3.common.atari_wrappers.AtariWrapper
   - - exploration_final_eps
@@ -10,20 +10,20 @@
   - - exploration_fraction
     - 0.1
   - - frame_stack
-    - 4
+    - 8
   - - gradient_steps
     - 1
   - - learning_rate
-    - 2.5e-05
+    - 0.0001
   - - learning_starts
-    - 100000
+    - 50000
   - - n_timesteps
-    - 1000000.0
+    - 200000.0
   - - optimize_memory_usage
     - false
   - - policy
     - CnnPolicy
   - - target_update_interval
-    - 1000
+    - 64
   - - train_freq
     - 32

dqn-SpaceInvadersNoFrameskip-v4.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f58adf1aa664f858520c84ad6ae193bd2352850fca5803945a01a3eee60ade54
-size 27220123
+oid sha256:3d9a7ada36ec248a9512dc3fd929b6a8164a1ebcf2fe69cf3681afde3e652076
+size 27541219

dqn-SpaceInvadersNoFrameskip-v4/policy.optimizer.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:347775f895140e7672be9e1447a679cad5f809e106db2af292aad609901dbaf8
-size 13505852
+oid sha256:301f4352b266f651b834b20c67d5644f1f1a112488640adc3c1a9806d5aff820
+size 13571388

dqn-SpaceInvadersNoFrameskip-v4/policy.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fe25a92409ef331c326674887542aaa6d0bf708bac0428c1b911d61cd5f80149
-size 13505178
+oid sha256:9d712ad4b8e1aed2dbe4921e45f459285596658cd5ceca460e4fc1390825f884
+size 13570714

dqn-SpaceInvadersNoFrameskip-v4/system_info.txt

@@ -1,7 +1,7 @@
 - OS: Linux-6.1.85+-x86_64-with-glibc2.35 # 1 SMP PREEMPT_DYNAMIC Thu Jun 27 21:05:47 UTC 2024
 - Python: 3.10.12
 - Stable-Baselines3: 2.4.0
-- PyTorch: 2.5.1+cpu
+- PyTorch: 2.5.1+cu121
 - GPU Enabled: False
 - Numpy: 1.26.4
 - Cloudpickle: 3.1.0

results.json

@@ -1 +1 @@
-{"mean_reward": 268.5, "std_reward": 78.16808811785025, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2024-12-03T19:33:29.554666"}
+{"mean_reward": 319.0, "std_reward": 105.9669759878048, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2024-12-04T16:02:19.019845"}

train_eval_metrics.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:009ae66109e75a5090bd285a22494d4dda50b0d95ee85064b48e77bf07632869
-size 5948
+oid sha256:63b973b3f8d5d5b48608a7671ce81cbdd956b0ebf252d1870e35706e6b34afc9
+size 5383