Initial commit

README.md

@@ -16,7 +16,7 @@ model-index:
       type: SpaceInvadersNoFrameskip-v4
     metrics:
     - type: mean_reward
-      value: 794.00 +/- 247.23
+      value: 713.50 +/- 271.45
       name: mean_reward
       verified: false
 ---
@@ -62,15 +62,15 @@ python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f lo
 
 ## Hyperparameters
 ```python
-OrderedDict([('batch_size', 84),
-             ('buffer_size', 100000),
+OrderedDict([('batch_size', 64),
+             ('buffer_size', 150000),
              ('env_wrapper',
               ['stable_baselines3.common.atari_wrappers.AtariWrapper']),
              ('exploration_final_eps', 0.01),
              ('exploration_fraction', 0.1),
              ('frame_stack', 4),
              ('gradient_steps', 1),
-             ('learning_rate', 0.001),
+             ('learning_rate', 5e-05),
              ('learning_starts', 100000),
              ('n_timesteps', 10000000.0),
              ('optimize_memory_usage', False),

args.yml

@@ -18,9 +18,9 @@
   - - gym_packages
     - []
   - - hyperparams
-    - batch_size: 84
-      buffer_size: 100000
-      learning_rate: 0.001
+    - batch_size: 64
+      buffer_size: 150000
+      learning_rate: 5.0e-05
   - - log_folder
     - logs/
   - - log_interval
@@ -36,7 +36,7 @@
   - - n_startup_trials
     - 10
   - - n_timesteps
-    - 100000
+    - 300000
   - - n_trials
     - 500
   - - no_optim_plots
@@ -58,7 +58,7 @@
   - - save_replay_buffer
     - false
   - - seed
-    - 274124340
+    - 1976347354
   - - storage
     - null
   - - study_name
@@ -68,7 +68,7 @@
   - - track
     - false
   - - trained_agent
-    - logs/dqn/SpaceInvadersNoFrameskip-v4_1/SpaceInvadersNoFrameskip-v4.zip
+    - logs/dqn/SpaceInvadersNoFrameskip-v4_14/best_model.zip
   - - truncate_last_trajectory
     - true
   - - uuid

config.yml

@@ -1,8 +1,8 @@
 !!python/object/apply:collections.OrderedDict
 - - - batch_size
-    - 84
+    - 64
   - - buffer_size
-    - 100000
+    - 150000
   - - env_wrapper
     - - stable_baselines3.common.atari_wrappers.AtariWrapper
   - - exploration_final_eps
@@ -14,7 +14,7 @@
   - - gradient_steps
     - 1
   - - learning_rate
-    - 0.001
+    - 5.0e-05
   - - learning_starts
     - 100000
   - - n_timesteps

dqn-SpaceInvadersNoFrameskip-v4.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd1650c5a90766289b63f09fe8645f73647ae46eb2256adb5157c107436faee9
+oid sha256:8f42db5926a09a365bab2a6bc2141a70662b2f71d4a43c392ce844f2f6ccf050
 size 27224804

dqn-SpaceInvadersNoFrameskip-v4/policy.optimizer.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a6edc7dc842a574379bede3a7fb4d41843b86641b9248d95b19cf959cc92aa0
+oid sha256:6dbbfac6b61c60d3ebba31ab991452ce57f398aed826b2b1b2a4c53672f86ef3
 size 13505739

dqn-SpaceInvadersNoFrameskip-v4/policy.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:007d3dd8a4fd60ba48ac4391bdb32a76b5c19153042cda913c91551ff594bcfb
+oid sha256:d48c41a27c1fc644c343e762042e7874c076aee5e1669f6099773707b06c7147
 size 13504937

replay.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e41ea531f1520c05a66684eec10c00f10517eb05cfa19ef4beda5810799ae6e
-size 239758
+oid sha256:10e6739dbc4fd2f9821608abf08818f6decdccd7a447b56fccf312dd8df5c683
+size 217644

results.json

@@ -1 +1 @@
-{"mean_reward": 794.0, "std_reward": 247.2326839234651, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2023-02-21T01:51:30.055701"}
+{"mean_reward": 713.5, "std_reward": 271.4502716889412, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2023-02-21T03:06:46.014286"}

train_eval_metrics.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:94d25b9d422cb8e0bf36ce3db93b5caa5950e032db31d26e3b41b84d60f85b7b
-size 5552
+oid sha256:b135317befaa660dfadb2e8e77f85efc03b7e730ff79e204c60a80aeb931bd5d
+size 10880