-
Notifications
You must be signed in to change notification settings - Fork 0
/
experiments.sh
executable file
·79 lines (66 loc) · 4.06 KB
/
experiments.sh
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
############# SIMPLE MADDPG #####################
# python run.py --model "maddpg" --env "simple" --experiment_name "continuous-v0" --total_memory 12 --episode_len 25 --batch_size 4 --n_episodes 10000 --num_layers 3 --num_filters 128 --communicate 0 --n_epochs 3
# python run.py --model "maddpg" --env "simple" --experiment_name "continuous-v1" --total_memory 12 --episode_len 25 --batch_size 4 --n_episodes 10000 --num_layers 3 --num_filters 128 --communicate 0 --n_epochs 3
# python run.py --model "maddpg" --env "simple" --experiment_name "continuous-v2" --total_memory 12 --episode_len 25 --batch_size 4 --n_episodes 10000 --num_layers 3 --num_filters 128 --communicate 0 --n_epochs 3
########################
# python run.py --model "maddpg" --env "spread" --experiment_name "v4-SquareDist" --total_memory 12 --episode_len 25 --batch_size 4 --n_episodes 25000 --num_layers 3 --num_filters 128 --communicate 0 --n_epochs 3
# python run.py --model "maddpg" --env "adversary" --experiment_name "v0" --total_memory 12 --episode_len 25 --batch_size 4 --n_episodes 40000 --num_layers 3 --num_filters 128 --communicate 0 --n_epochs 3
# python run.py --model "ppo_rnn_policy_shared" --env "simple" --experiment_name "ppo-rnn" --episode_len 25 --num-envs 8
# python run.py --model "ppo_rec_global_critic" --env "full_communication_3" --experiment_name "main" --num-envs 512 --total-timesteps 75000 --learning-rate 0.0007 --update-epochs 10 --max-grad-norm 10 --episode_len 25 --wandb True #--load_weights_name "/ppo_rec_global_critic-full_communication_2-main"
# python run.py --model "ppo_no_scaling_rec_global_critic" --env "full_communication_2" --experiment_name "main" --num-envs 256 --total-timesteps 25000 --learning-rate 0.0007 --update-epochs 10 --max-grad-norm 10 --episode_len 25 --wandb True #--load_weights_name "/ppo_rec_global_critic-full_communication_2-main"
# python run.py --model "ppo_no_scaling_rec_global_critic" --env "full_communication_3" --experiment_name "no_init_share_representation" --num-envs 512 --total-timesteps 75000 --learning-rate 0.0007 --update-epochs 10 --max-grad-norm 10 --episode_len 25 --wandb True #--load_weights_name "/ppo_no_scaling_rec_global_critic-full_communication_2-main"
# python run.py \
# --model "ppo_shared_global_critic_rec" \
# --env "full_communication_3" \
# --experiment_name "report_env" \
# --num-envs 1024 \
# --total-timesteps 50000 \
# --learning-rate 0.0007 \
# --update-epochs 10 \
# --max-grad-norm 10 \
# --episode_len 25 \
# --wandb True \
# --video False
# python run.py \
# --model "ppo_shared_use_future" \
# --env "full_communication_2" \
# --experiment_name "big_net_bigger_bottleneck" \
# --total-episodes 1000000 \
# --learning-rate 0.0007 \
# --batch_size 512 \
# --update-epochs 10 \
# --max-grad-norm 10 \
# --episode_len 25 \
# --hidden_size 128 \
# --wandb False \
# --video True \
# --load_weights_name "/ppo_shared_use_future-full_communication_4-big_net_cont"
python iterated_run.py \
--model "language_learner_agent" \
--env "iterated" \
--experiment_name "FINAL_ITERATED" \
--total-episodes 2500000 \
--learning-rate 0.0007 \
--batch_size 512 \
--update-epochs 10 \
--max-grad-norm 10 \
--episode_len 25 \
--hidden_size 256 \
--wandb True \
--video True \
# --load_weights_name "/language_learner_agent-iterated-iterated_2"
# python run.py \
# --model "ppo_shared_global_critic_rec" \
# --env "full_communication_4" \
# --experiment_name "sum_com" \
# --num-envs 256 \
# --total-timesteps 100000 \
# --learning-rate 0.0007 \
# --update-epochs 10 \
# --max-grad-norm 10 \
# --episode_len 25 \
# --wandb True \
# --video True \
# --load_weights_name "/ppo_shared_global_critic_rec-full_communication_3-sum_com"
# python run.py --model "ppo_policy3" --env "simple" --experiment_name "test" --episode_len 25 --num-envs 8
# python run.py --model "ppo_policy3_shared" --env "communication_full" --experiment_name "reference_shared_info" --episode_len 25 --num-envs 8