bacow/easy-rl
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

update rainbowdqn

Browse Source
This commit is contained in:
johnjim0816
2022-05-31 01:20:58 +08:00
parent cfc0f6492e
commit c7c94468c9
149 changed files with 1866 additions and 1549 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
codes/SoftActorCritic/model.py
View File

@@ -17,10 +17,10 @@ from torch.distributions import Normal
device=torch.device("cuda" if torch.cuda.is_available() else "cpu")
class ValueNet(nn.Module):
def __init__(self, state_dim, hidden_dim, init_w=3e-3):
def __init__(self, n_states, hidden_dim, init_w=3e-3):
super(ValueNet, self).__init__()
self.linear1 = nn.Linear(state_dim, hidden_dim)
self.linear1 = nn.Linear(n_states, hidden_dim)
self.linear2 = nn.Linear(hidden_dim, hidden_dim)
self.linear3 = nn.Linear(hidden_dim, 1)
@@ -35,10 +35,10 @@ class ValueNet(nn.Module):
class SoftQNet(nn.Module):
def __init__(self, state_dim, action_dim, hidden_dim, init_w=3e-3):
def __init__(self, n_states, n_actions, hidden_dim, init_w=3e-3):
super(SoftQNet, self).__init__()
self.linear1 = nn.Linear(state_dim + action_dim, hidden_dim)
self.linear1 = nn.Linear(n_states + n_actions, hidden_dim)
self.linear2 = nn.Linear(hidden_dim, hidden_dim)
self.linear3 = nn.Linear(hidden_dim, 1)
@@ -54,20 +54,20 @@ class SoftQNet(nn.Module):
class PolicyNet(nn.Module):
def __init__(self, state_dim, action_dim, hidden_dim, init_w=3e-3, log_std_min=-20, log_std_max=2):
def __init__(self, n_states, n_actions, hidden_dim, init_w=3e-3, log_std_min=-20, log_std_max=2):
super(PolicyNet, self).__init__()
self.log_std_min = log_std_min
self.log_std_max = log_std_max
self.linear1 = nn.Linear(state_dim, hidden_dim)
self.linear1 = nn.Linear(n_states, hidden_dim)
self.linear2 = nn.Linear(hidden_dim, hidden_dim)
self.mean_linear = nn.Linear(hidden_dim, action_dim)
self.mean_linear = nn.Linear(hidden_dim, n_actions)
self.mean_linear.weight.data.uniform_(-init_w, init_w)
self.mean_linear.bias.data.uniform_(-init_w, init_w)
self.log_std_linear = nn.Linear(hidden_dim, action_dim)
self.log_std_linear = nn.Linear(hidden_dim, n_actions)
self.log_std_linear.weight.data.uniform_(-init_w, init_w)
self.log_std_linear.bias.data.uniform_(-init_w, init_w)

26
codes/SoftActorCritic/sac.py
View File

@@ -43,10 +43,10 @@ class ReplayBuffer:
return len(self.buffer)
class ValueNet(nn.Module):
def __init__(self, state_dim, hidden_dim, init_w=3e-3):
def __init__(self, n_states, hidden_dim, init_w=3e-3):
super(ValueNet, self).__init__()
self.linear1 = nn.Linear(state_dim, hidden_dim)
self.linear1 = nn.Linear(n_states, hidden_dim)
self.linear2 = nn.Linear(hidden_dim, hidden_dim)
self.linear3 = nn.Linear(hidden_dim, 1)
@@ -61,10 +61,10 @@ class ValueNet(nn.Module):
class SoftQNet(nn.Module):
def __init__(self, state_dim, action_dim, hidden_dim, init_w=3e-3):
def __init__(self, n_states, n_actions, hidden_dim, init_w=3e-3):
super(SoftQNet, self).__init__()
self.linear1 = nn.Linear(state_dim + action_dim, hidden_dim)
self.linear1 = nn.Linear(n_states + n_actions, hidden_dim)
self.linear2 = nn.Linear(hidden_dim, hidden_dim)
self.linear3 = nn.Linear(hidden_dim, 1)
@@ -80,20 +80,20 @@ class SoftQNet(nn.Module):
class PolicyNet(nn.Module):
def __init__(self, state_dim, action_dim, hidden_dim, init_w=3e-3, log_std_min=-20, log_std_max=2):
def __init__(self, n_states, n_actions, hidden_dim, init_w=3e-3, log_std_min=-20, log_std_max=2):
super(PolicyNet, self).__init__()
self.log_std_min = log_std_min
self.log_std_max = log_std_max
self.linear1 = nn.Linear(state_dim, hidden_dim)
self.linear1 = nn.Linear(n_states, hidden_dim)
self.linear2 = nn.Linear(hidden_dim, hidden_dim)
self.mean_linear = nn.Linear(hidden_dim, action_dim)
self.mean_linear = nn.Linear(hidden_dim, n_actions)
self.mean_linear.weight.data.uniform_(-init_w, init_w)
self.mean_linear.bias.data.uniform_(-init_w, init_w)
self.log_std_linear = nn.Linear(hidden_dim, action_dim)
self.log_std_linear = nn.Linear(hidden_dim, n_actions)
self.log_std_linear.weight.data.uniform_(-init_w, init_w)
self.log_std_linear.bias.data.uniform_(-init_w, init_w)
@@ -134,14 +134,14 @@ class PolicyNet(nn.Module):
return action[0]
class SAC:
def __init__(self,state_dim,action_dim,cfg) -> None:
def __init__(self,n_states,n_actions,cfg) -> None:
self.batch_size = cfg.batch_size
self.memory = ReplayBuffer(cfg.capacity)
self.device = cfg.device
self.value_net = ValueNet(state_dim, cfg.hidden_dim).to(self.device)
self.target_value_net = ValueNet(state_dim, cfg.hidden_dim).to(self.device)
self.soft_q_net = SoftQNet(state_dim, action_dim, cfg.hidden_dim).to(self.device)
self.policy_net = PolicyNet(state_dim, action_dim, cfg.hidden_dim).to(self.device)
self.value_net = ValueNet(n_states, cfg.hidden_dim).to(self.device)
self.target_value_net = ValueNet(n_states, cfg.hidden_dim).to(self.device)
self.soft_q_net = SoftQNet(n_states, n_actions, cfg.hidden_dim).to(self.device)
self.policy_net = PolicyNet(n_states, n_actions, cfg.hidden_dim).to(self.device)
self.value_optimizer = optim.Adam(self.value_net.parameters(), lr=cfg.value_lr)
self.soft_q_optimizer = optim.Adam(self.soft_q_net.parameters(), lr=cfg.soft_q_lr)
self.policy_optimizer = optim.Adam(self.policy_net.parameters(), lr=cfg.policy_lr)

6
codes/SoftActorCritic/task0.py
View File

@@ -63,9 +63,9 @@ class PlotConfig:
def env_agent_config(cfg,seed=1):
env = NormalizedActions(gym.make(cfg.env_name))
env.seed(seed)
action_dim = env.action_space.shape[0]
state_dim = env.observation_space.shape[0]
agent = SAC(state_dim,action_dim,cfg)
n_actions = env.action_space.shape[0]
n_states = env.observation_space.shape[0]
agent = SAC(n_states,n_actions,cfg)
return env,agent
def train(cfg,env,agent):

8
codes/SoftActorCritic/task0_train.ipynb
View File

@@ -70,9 +70,9 @@
"def env_agent_config(cfg,seed=1):\n",
" env = NormalizedActions(gym.make(\"Pendulum-v0\"))\n",
" env.seed(seed)\n",
" action_dim = env.action_space.shape[0]\n",
" state_dim = env.observation_space.shape[0]\n",
" agent = SAC(state_dim,action_dim,cfg)\n",
" n_actions = env.action_space.shape[0]\n",
" n_states = env.observation_space.shape[0]\n",
" agent = SAC(n_states,n_actions,cfg)\n",
" return env,agent"
]
},
@@ -159,7 +159,7 @@
"\nDuring handling of the above exception, another exception occurred:\n",
"\u001b[0;31mDeprecatedEnv\u001b[0m Traceback (most recent call last)",
"\u001b[0;32m<ipython-input-7-91b1038013e4>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m 3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 4\u001b[0m \u001b[0;31m# train\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0menv\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0magent\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0menv_agent_config\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcfg\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mseed\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 6\u001b[0m \u001b[0mrewards\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mma_rewards\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtrain\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcfg\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0menv\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0magent\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 7\u001b[0m \u001b[0mmake_dir\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcfg\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mresult_path\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcfg\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodel_path\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m<ipython-input-4-040773221550>\u001b[0m in \u001b[0;36menv_agent_config\u001b[0;34m(cfg, seed)\u001b[0m\n\u001b[1;32m 1\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0menv_agent_config\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcfg\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mseed\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0menv\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mNormalizedActions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgym\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmake\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Pendulum-v0\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 3\u001b[0m \u001b[0menv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mseed\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mseed\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 4\u001b[0m \u001b[0maction_dim\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0menv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0maction_space\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 5\u001b[0m \u001b[0mstate_dim\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0menv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mobservation_space\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m<ipython-input-4-040773221550>\u001b[0m in \u001b[0;36menv_agent_config\u001b[0;34m(cfg, seed)\u001b[0m\n\u001b[1;32m 1\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0menv_agent_config\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcfg\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mseed\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0menv\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mNormalizedActions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgym\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmake\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Pendulum-v0\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 3\u001b[0m \u001b[0menv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mseed\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mseed\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 4\u001b[0m \u001b[0mn_actions\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0menv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0maction_space\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 5\u001b[0m \u001b[0mn_states\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0menv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mobservation_space\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m~/anaconda3/envs/py37/lib/python3.7/site-packages/gym/envs/registration.py\u001b[0m in \u001b[0;36mmake\u001b[0;34m(id, **kwargs)\u001b[0m\n\u001b[1;32m 233\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 234\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mmake\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mid\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 235\u001b[0;31m \u001b[0;32mreturn\u001b[0m \u001b[0mregistry\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmake\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mid\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 236\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 237\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m~/anaconda3/envs/py37/lib/python3.7/site-packages/gym/envs/registration.py\u001b[0m in \u001b[0;36mmake\u001b[0;34m(self, path, **kwargs)\u001b[0m\n\u001b[1;32m 126\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 127\u001b[0m \u001b[0mlogger\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minfo\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Making new env: %s\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 128\u001b[0;31m \u001b[0mspec\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mspec\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 129\u001b[0m \u001b[0menv\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mspec\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmake\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 130\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0menv\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m~/anaconda3/envs/py37/lib/python3.7/site-packages/gym/envs/registration.py\u001b[0m in \u001b[0;36mspec\u001b[0;34m(self, path)\u001b[0m\n\u001b[1;32m 185\u001b[0m raise error.DeprecatedEnv(\n\u001b[1;32m 186\u001b[0m \"Env {} not found (valid versions include {})\".format(\n\u001b[0;32m--> 187\u001b[0;31m \u001b[0mid\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmatching_envs\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 188\u001b[0m )\n\u001b[1;32m 189\u001b[0m )\n",