Hugging Face's logo

pfunk
/
CartPole-v1-CP_DQPN_x1-seed4

Reinforcement Learning
TensorBoard
cleanrl
CartPole-v1
deep-reinforcement-learning
custom-implementation
Eval Results
Model card Files Metrics Community
mmorales34 commited on
Commit
5afd521
•
1 Parent(s): 35182c6

pushing model

Browse files
Files changed (5) hide show
  1. dqpn_freq.py +3 -4
  2. events.out.tfevents.1679277764.rhea.184065.0 → events.out.tfevents.1679289421.redi.466045.0 +1 -1
  3. videos/{CartPole-v1__CP_DQPN_x1__4__1679277760-eval → CartPole-v1__CP_DQPN_x1__4__1679289417-eval}/rl-video-episode-0.mp4 +0 -0
  4. videos/{CartPole-v1__CP_DQPN_x1__4__1679277760-eval → CartPole-v1__CP_DQPN_x1__4__1679289417-eval}/rl-video-episode-1.mp4 +0 -0
  5. videos/{CartPole-v1__CP_DQPN_x1__4__1679277760-eval → CartPole-v1__CP_DQPN_x1__4__1679289417-eval}/rl-video-episode-8.mp4 +0 -0
dqpn_freq.py CHANGED
@@ -141,17 +141,16 @@ if __name__ == "__main__":
141
  wandb.log({name: x, "global_step": y})
142
 
143
  # TRY NOT TO MODIFY: seeding
144
- random.seed(args.seed)
145
- np.random.seed(args.seed)
146
  torch.manual_seed(args.seed)
147
- torch.backends.cudnn.deterministic = args.torch_deterministic
 
 
148
 
149
  device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
150
 
151
  # env setup
152
  envs = gym.vector.SyncVectorEnv([make_env(args.env_id, args.seed, 0, args.capture_video, run_name)])
153
  assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
154
- envs.seed(args.seed)
155
 
156
  q_network = QNetwork(envs).to(device)
157
  optimizer = optim.RMSprop(q_network.parameters(), lr=args.learning_rate)
 
141
  wandb.log({name: x, "global_step": y})
142
 
143
  # TRY NOT TO MODIFY: seeding
 
 
144
  torch.manual_seed(args.seed)
145
+ # torch.backends.cudnn.deterministic = args.torch_deterministic
146
+ np.random.seed(args.seed)
147
+ random.seed(args.seed)
148
 
149
  device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
150
 
151
  # env setup
152
  envs = gym.vector.SyncVectorEnv([make_env(args.env_id, args.seed, 0, args.capture_video, run_name)])
153
  assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
 
154
 
155
  q_network = QNetwork(envs).to(device)
156
  optimizer = optim.RMSprop(q_network.parameters(), lr=args.learning_rate)
events.out.tfevents.1679277764.rhea.184065.0 → events.out.tfevents.1679289421.redi.466045.0 RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bdb42a82e704fb799b373d244a84fb4ab36f05e95a58160569ce475f30d9f18e
3
  size 684
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:364b77c2d13f403386deb8dab1d8cd9480e433e437b76a3f8be620b8911463eb
3
  size 684
videos/{CartPole-v1__CP_DQPN_x1__4__1679277760-eval → CartPole-v1__CP_DQPN_x1__4__1679289417-eval}/rl-video-episode-0.mp4 RENAMED
File without changes
videos/{CartPole-v1__CP_DQPN_x1__4__1679277760-eval → CartPole-v1__CP_DQPN_x1__4__1679289417-eval}/rl-video-episode-1.mp4 RENAMED
File without changes
videos/{CartPole-v1__CP_DQPN_x1__4__1679277760-eval → CartPole-v1__CP_DQPN_x1__4__1679289417-eval}/rl-video-episode-8.mp4 RENAMED
File without changes