pushing model

.gitattributes

@@ -42,3 +42,4 @@ videos/Walker2d-v4__ppo_fix_continuous_action__5__1704458010-eval/rl-video-episo
 videos/Walker2d-v4__ppo_fix_continuous_action__5__1704498502-eval/rl-video-episode-8.mp4 filter=lfs diff=lfs merge=lfs -text
 videos/Walker2d-v4__ppo_fix_continuous_action__5__1704498502-eval/rl-video-episode-1.mp4 filter=lfs diff=lfs merge=lfs -text
 videos/Walker2d-v4__ppo_fix_continuous_action__5__1704498502-eval/rl-video-episode-0.mp4 filter=lfs diff=lfs merge=lfs -text
+videos/Walker2d-v4__ppo_fix_continuous_action__5__1705721020-eval/rl-video-episode-0.mp4 filter=lfs diff=lfs merge=lfs -text

README.md

@@ -16,7 +16,7 @@ model-index:
       type: Walker2d-v4
     metrics:
     - type: mean_reward
-      value: 4924.31 +/- 1237.76
+      value: 1457.49 +/- 811.87
       name: mean_reward
       verified: false
 ---
@@ -46,7 +46,7 @@ curl -OL https://huggingface.co/sdpkjc/Walker2d-v4-ppo_fix_continuous_action-see
 curl -OL https://huggingface.co/sdpkjc/Walker2d-v4-ppo_fix_continuous_action-seed5/raw/main/pyproject.toml
 curl -OL https://huggingface.co/sdpkjc/Walker2d-v4-ppo_fix_continuous_action-seed5/raw/main/poetry.lock
 poetry install --all-extras
-python ppo_fix_continuous_action.py --save-model --upload-model --hf-entity sdpkjc --env-id Walker2d-v4 --seed 5 --track
+python ppo_fix_continuous_action.py --save-model --upload-model --hf-entity sdpkjc --env-id Walker2d-v4 --seed 5 --track --cuda False
 ```
 
 # Hyperparameters
@@ -56,7 +56,7 @@ python ppo_fix_continuous_action.py --save-model --upload-model --hf-entity sdpk
  'capture_video': False,
  'clip_coef': 0.2,
  'clip_vloss': True,
- 'cuda': True,
+ 'cuda': False,
  'ent_coef': 0.0,
  'env_id': 'Walker2d-v4',
  'exp_name': 'ppo_fix_continuous_action',

events.out.tfevents.1704498509.4090-171.326604.0 → events.out.tfevents.1705721030.3090-172.2802263.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bf8dfc6862d55e236cdfdea7ee7b32ffb04e004e10f4f14bd13785b4d2f25ecc
-size 573863
+oid sha256:6279a7e79abe1861119b9dd2611ab27766decc58ba4c1098c0a8b49a29c29239
+size 791356

ppo_fix_continuous_action.py

@@ -198,7 +198,7 @@ class NormalizeReward(gym.core.Wrapper, gym.utils.RecordConstructorArgs):
         return obs, rews, terminateds, truncateds, infos
 
     def reset(self, **kwargs):
-        self.returns = np.zeros(self.num_envs)
+        # self.returns = np.zeros(self.num_envs)
         return self.env.reset(**kwargs)
 
     def normalize(self, rews):

replay.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6c538c3c9d4183120e0678895787edb25663ab5d14c8b2e3593fc9853735c5d9
-size 1334268
+oid sha256:8925ae3b805d55df0e466a055ad06056ffcbbe556795b353340370247e6fb683
+size 279554

videos/Walker2d-v4__ppo_fix_continuous_action__5__1704498502-eval/rl-video-episode-1.mp4

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:20ed525788b193b6b795e5a60660b590d6e6ee388767822566aee42260186f8f
-size 1335012

videos/Walker2d-v4__ppo_fix_continuous_action__5__1704498502-eval/rl-video-episode-8.mp4

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6c538c3c9d4183120e0678895787edb25663ab5d14c8b2e3593fc9853735c5d9
-size 1334268

videos/{Walker2d-v4__ppo_fix_continuous_action__5__1704498502-eval → Walker2d-v4__ppo_fix_continuous_action__5__1705721020-eval}/rl-video-episode-0.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a6e6bae41f224bb8f45d64aa9675adb7d0348bb55897e7db23297c65ca03627
-size 1312361
+oid sha256:690cb6a7c92c13cc9bd6b6c4934e39e0fb024837609dfa6ea9ca0c7b73cd40f4
+size 1325646