Initial commit

README.md

@@ -16,7 +16,7 @@ model-index:
       type: SpaceInvadersNoFrameskip-v4
     metrics:
     - type: mean_reward
-      value: 29.00 +/- 64.30
+      value: 617.00 +/- 116.11
       name: mean_reward
       verified: false
 ---
@@ -62,8 +62,8 @@ python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f lo
 
 ## Hyperparameters
 ```python
-OrderedDict([('batch_size', 32),
-             ('buffer_size', 100000),
+OrderedDict([('batch_size', 64),
+             ('buffer_size', 250000),
              ('env_wrapper',
               ['stable_baselines3.common.atari_wrappers.AtariWrapper']),
              ('exploration_final_eps', 0.01),
@@ -72,10 +72,10 @@ OrderedDict([('batch_size', 32),
              ('gradient_steps', 1),
              ('learning_rate', 0.0001),
              ('learning_starts', 100000),
-             ('n_timesteps', 100000.0),
+             ('n_timesteps', 2000000.0),
              ('optimize_memory_usage', False),
              ('policy', 'CnnPolicy'),
-             ('target_update_interval', 1000),
+             ('target_update_interval', 900),
              ('train_freq', 4),
              ('normalize', False)])
 ```

args.yml

@@ -54,7 +54,7 @@
   - - save_replay_buffer
     - false
   - - seed
-    - 2326548282
+    - 380237253
   - - storage
     - null
   - - study_name

config.yml

@@ -1,8 +1,8 @@
 !!python/object/apply:collections.OrderedDict
 - - - batch_size
-    - 32
+    - 64
   - - buffer_size
-    - 100000
+    - 250000
   - - env_wrapper
     - - stable_baselines3.common.atari_wrappers.AtariWrapper
   - - exploration_final_eps
@@ -18,12 +18,12 @@
   - - learning_starts
     - 100000
   - - n_timesteps
-    - 100000.0
+    - 2000000.0
   - - optimize_memory_usage
     - false
   - - policy
     - CnnPolicy
   - - target_update_interval
-    - 1000
+    - 900
   - - train_freq
     - 4

dqn-SpaceInvadersNoFrameskip-v4.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b76a480fcc48fe2bdf40060d849e6a26c3cf6f7d91d3f11ea4dc892380712785
-size 13715083
+oid sha256:e4d6167e887f6be808e766d8f67103e7beec42a7968f32949436530c0185a58c
+size 27220203

dqn-SpaceInvadersNoFrameskip-v4/_stable_baselines3_version

@@ -1 +1 @@
-2.3.0a2
+2.3.0

dqn-SpaceInvadersNoFrameskip-v4/policy.optimizer.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e3fc84919e676e2bdcaa04814c7dfa2b7625d8d53a818418c6db25771364db93
-size 1120
+oid sha256:0ca7e2ff48bd15dd01427090c13dd6f2232a3e6271ace0b3da61eb64dac68109
+size 13506236

dqn-SpaceInvadersNoFrameskip-v4/policy.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bca1a7e1a6744fcec5d18ffc1f5748b491d62cf7631935e3f8bbcd2d07bf0045
+oid sha256:42b2ecf5c42f3c3ebb150f312d152118d4ac02053dd02a5b92cfd73b462618e8
 size 13505370

dqn-SpaceInvadersNoFrameskip-v4/system_info.txt

@@ -1,7 +1,7 @@
 - OS: Linux-6.1.58+-x86_64-with-glibc2.35 # 1 SMP PREEMPT_DYNAMIC Sat Nov 18 15:31:17 UTC 2023
 - Python: 3.10.12
-- Stable-Baselines3: 2.3.0a2
-- PyTorch: 2.1.0+cu121
+- Stable-Baselines3: 2.3.0
+- PyTorch: 2.2.1+cu121
 - GPU Enabled: True
 - Numpy: 1.25.2
 - Cloudpickle: 2.2.1

replay.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d8dc44ae099d33c1b445ec14463db23469bb584dcc8e1918c1d1d797485e127f
-size 294856
+oid sha256:743ae2973b1443e786c88516b4c3b98f0f0827f91154bfbc519e95eac592995b
+size 235375

results.json

@@ -1 +1 @@
-{"mean_reward": 29.0, "std_reward": 64.29618962271404, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2024-02-19T20:15:27.178591"}
+{"mean_reward": 617.0, "std_reward": 116.1077086157504, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2024-03-31T22:51:31.793692"}

train_eval_metrics.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2278e9e44bbc1dd331febbccd4d503d8ed7f562c1ae6785a856e83ac644d6430
-size 5619
+oid sha256:ffcb6f8f4c4d46699dcaba1b691f4379242bd75fc9d218684fba81cee989d4ba
+size 61647