Initial commit

README.md

@@ -16,7 +16,7 @@ model-index:
       type: SpaceInvadersNoFrameskip-v4
     metrics:
     - type: mean_reward
-      value: 329.00 +/- 157.97
+      value: 257.00 +/- 38.81
       name: mean_reward
       verified: false
 ---
@@ -62,7 +62,7 @@ python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f lo
 
 ## Hyperparameters
 ```python
-OrderedDict([('batch_size', 16),
+OrderedDict([('batch_size', 64),
              ('buffer_size', 150000),
              ('env_wrapper',
               ['stable_baselines3.common.atari_wrappers.AtariWrapper']),
@@ -70,9 +70,9 @@ OrderedDict([('batch_size', 16),
              ('exploration_fraction', 0.1),
              ('frame_stack', 4),
              ('gradient_steps', 1),
-             ('learning_rate', 2e-05),
+             ('learning_rate', 1e-06),
              ('learning_starts', 100000),
-             ('n_timesteps', 10000.0),
+             ('n_timesteps', 90000.0),
              ('optimize_memory_usage', False),
              ('policy', 'CnnPolicy'),
              ('target_update_interval', 1000),

args.yml

@@ -54,7 +54,7 @@
   - - save_replay_buffer
     - false
   - - seed
-    - 942847196
+    - 798114142
   - - storage
     - null
   - - study_name

config.yml

@@ -1,6 +1,6 @@
 !!python/object/apply:collections.OrderedDict
 - - - batch_size
-    - 16
+    - 64
   - - buffer_size
     - 150000
   - - env_wrapper
@@ -14,11 +14,11 @@
   - - gradient_steps
     - 1
   - - learning_rate
-    - 2.0e-05
+    - 1.0e-06
   - - learning_starts
     - 100000
   - - n_timesteps
-    - 10000.0
+    - 90000.0
   - - optimize_memory_usage
     - false
   - - policy

dqn-SpaceInvadersNoFrameskip-v4.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:545447b52726e403bcce1b0f5a38c35718425185f1641aea1b005c33ef10c59c
-size 13711811
+oid sha256:6c965789154b1a43487e978eb2a8647a2943a0b39ea9408977b253e01ad7f025
+size 13715080

dqn-SpaceInvadersNoFrameskip-v4/policy.optimizer.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0cc351647a59bfe65acdd23d06b233cc39b75fad5d3d7450f37fb5322659f8d2
+oid sha256:c1f3b836bd5d42486a268ff007163ab7fddb9054f9656d4a7efaad53c15e95a8
 size 1120

dqn-SpaceInvadersNoFrameskip-v4/policy.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:caf52418779159e9a18f9eabfe9cacbdb8c4909e7c30119e5c5e265578ec94cc
+oid sha256:320fc5512d555634dd6c38369d2616ed265c5b9a22a26222370711079503f726
 size 13505370

replay.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d1220fd2663a829192a2b709c767eb2debf32246e6f8d3b5838a5840975ff8d0
-size 242623
+oid sha256:961ab6016871f866ed78c92d21ef7930dec0110e72144703e0be2b439e42e1af
+size 267942

results.json

@@ -1 +1 @@
-{"mean_reward": 329.0, "std_reward": 157.96835126062436, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2024-03-11T22:04:31.141072"}
+{"mean_reward": 257.0, "std_reward": 38.80721582386451, "is_deterministic": false, "n_eval_episodes": 10, "eval_datetime": "2024-03-11T22:15:05.151872"}

train_eval_metrics.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c922f99b683ff977decf29898d5df84f4c9d440b6eff971d94fbbcbc559cf0a4
-size 646
+oid sha256:e3a9a348166e055a970c9f6bfdddcb2ad18e036c179937e58912383c747049c6
+size 5268