Hugging Face's logo

Spaces:
gsc579
/
joyrl-benckmarks
No application file

App Files Community
gsc579 commited on
Commit
2a12873
•
1 Parent(s): f7c8dad

add ppo-hopper-v4

Browse files
Files changed (6) hide show
  1. Train_single_Hopper-v4_PPO_20230526-135633/config.yaml +73 -0
  2. Train_single_Hopper-v4_PPO_20230526-135633/logs/log.txt +0 -0
  3. Train_single_Hopper-v4_PPO_20230526-135633/models/500 +0 -0
  4. Train_single_Hopper-v4_PPO_20230526-135633/models/best +0 -0
  5. Train_single_Hopper-v4_PPO_20230526-135633/tb_logs/interact/events.out.tfevents.1685109393.gscaimax.2069774.0 +3 -0
  6. Train_single_Hopper-v4_PPO_20230526-135633/tb_logs/model/events.out.tfevents.1685109393.gscaimax.2069774.1 +3 -0
Train_single_Hopper-v4_PPO_20230526-135633/config.yaml ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ general_cfg:
2
+ algo_name: PPO
3
+ collect_traj: false
4
+ device: cuda
5
+ env_name: gym
6
+ load_checkpoint: false
7
+ load_model_step: best
8
+ load_path: Train_single_CartPole-v1_DQN_20230515-211721
9
+ max_episode: 2000
10
+ max_step: 1000
11
+ mode: train
12
+ model_save_fre: 500
13
+ mp_backend: single
14
+ n_learners: 1
15
+ n_workers: 2
16
+ online_eval: true
17
+ online_eval_episode: 10
18
+ seed: 1
19
+ share_buffer: true
20
+ algo_cfg:
21
+ actor_hidden_dim: 256
22
+ actor_layers:
23
+ - activation: relu
24
+ layer_size:
25
+ - 256
26
+ layer_type: linear
27
+ - activation: relu
28
+ layer_size:
29
+ - 256
30
+ layer_type: linear
31
+ actor_lr: 0.0003
32
+ batch_size: 256
33
+ buffer_size: 100000
34
+ buffer_type: ONPOLICY_QUE
35
+ continuous: true
36
+ critic_hidden_dim: 256
37
+ critic_layers:
38
+ - activation: relu
39
+ layer_size:
40
+ - 256
41
+ layer_type: linear
42
+ - activation: relu
43
+ layer_size:
44
+ - 256
45
+ layer_type: linear
46
+ critic_loss_coef: 0.5
47
+ critic_lr: 0.001
48
+ entropy_coef: 0.01
49
+ eps_clip: 0.2
50
+ epsilon_decay: 500
51
+ epsilon_end: 0.01
52
+ epsilon_start: 0.95
53
+ gamma: 0.95
54
+ independ_actor: true
55
+ k_epochs: 8
56
+ kl_alpha: 2
57
+ kl_beta: 1.5
58
+ kl_lambda: 0.5
59
+ kl_target: 0.1
60
+ lr: 0.0001
61
+ min_policy: 0
62
+ ppo_type: clip
63
+ sgd_batch_size: 128
64
+ share_optimizer: false
65
+ target_update: 4
66
+ env_cfg:
67
+ id: Hopper-v4
68
+ ignore_params:
69
+ - wrapper
70
+ - ignore_params
71
+ new_step_api: true
72
+ render_mode: null
73
+ wrapper: null
Train_single_Hopper-v4_PPO_20230526-135633/logs/log.txt ADDED
The diff for this file is too large to render. See raw diff
 
Train_single_Hopper-v4_PPO_20230526-135633/models/500 ADDED
Binary file (559 kB). View file
 
Train_single_Hopper-v4_PPO_20230526-135633/models/best ADDED
Binary file (559 kB). View file
 
Train_single_Hopper-v4_PPO_20230526-135633/tb_logs/interact/events.out.tfevents.1685109393.gscaimax.2069774.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d336f8ce6688cc086c520124c6229dfb70d2c7cb4809fbb4d34c1980b0bce0f7
3
+ size 211780
Train_single_Hopper-v4_PPO_20230526-135633/tb_logs/model/events.out.tfevents.1685109393.gscaimax.2069774.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92162a641de15d129708d2ca49d12766fcff60b186e93d4cecb59f86b89d2176
3
+ size 129875