-
Notifications
You must be signed in to change notification settings - Fork 5
/
Copy pathtrain.py
executable file
·149 lines (139 loc) · 6.6 KB
/
train.py
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
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
"""
@author: Viet Nguyen <[email protected]>
"""
import os
os.environ['OMP_NUM_THREADS'] = '1'
import argparse
import torch
from src.env import MultipleEnvironments
from src.model import PPO
from src.process import test
import torch.multiprocessing as _mp
from torch.distributions import Categorical
import torch.nn.functional as F
import numpy as np
import shutil
def get_args():
parser = argparse.ArgumentParser(
"""Implementation of model described in the paper: Proximal Policy Optimization Algorithms for Sonic The Hedgehog""")
parser.add_argument("--zone", type=int, default=1)
parser.add_argument("--act", type=int, default=1)
parser.add_argument('--lr', type=float, default=1e-4)
parser.add_argument('--gamma', type=float, default=0.99, help='discount factor for rewards')
parser.add_argument('--tau', type=float, default=1.0, help='parameter for GAE')
parser.add_argument('--beta', type=float, default=0.01, help='entropy coefficient')
parser.add_argument('--epsilon', type=float, default=0.2, help='parameter for Clipped Surrogate Objective')
parser.add_argument('--batch_size', type=int, default=8)
parser.add_argument('--num_epochs', type=int, default=10)
parser.add_argument("--num_local_steps", type=int, default=512)
parser.add_argument("--num_global_steps", type=int, default=5e6)
parser.add_argument("--num_processes", type=int, default=12)
parser.add_argument("--save_interval", type=int, default=200, help="Number of steps between savings")
parser.add_argument("--max_actions", type=int, default=500, help="Maximum repetition steps in test phase")
parser.add_argument("--log_path", type=str, default="tensorboard/ppo_SonicTheHedgehog")
parser.add_argument("--saved_path", type=str, default="trained_models")
parser.add_argument("--load_from_previous_stage", type=bool, default=False,
help="Load weight from previous trained stage")
args = parser.parse_args()
return args
def train(opt):
if torch.cuda.is_available():
torch.cuda.manual_seed(123)
else:
torch.manual_seed(123)
if os.path.isdir(opt.log_path):
shutil.rmtree(opt.log_path)
os.makedirs(opt.log_path)
if not os.path.isdir(opt.saved_path):
os.makedirs(opt.saved_path)
mp = _mp.get_context("spawn")
envs = MultipleEnvironments(opt.zone, opt.act, opt.num_processes)
model = PPO(envs.num_states, envs.num_actions)
if torch.cuda.is_available():
model.cuda()
model.share_memory()
process = mp.Process(target=test, args=(opt, model, envs.num_states, envs.num_actions))
process.start()
optimizer = torch.optim.Adam(model.parameters(), lr=opt.lr)
[agent_conn.send(("reset", None)) for agent_conn in envs.agent_conns]
curr_states = [agent_conn.recv() for agent_conn in envs.agent_conns]
curr_states = torch.from_numpy(np.concatenate(curr_states, 0))
if torch.cuda.is_available():
curr_states = curr_states.cuda()
curr_episode = 0
while True:
curr_episode += 1
old_log_policies = []
actions = []
values = []
states = []
rewards = []
dones = []
for _ in range(opt.num_local_steps):
states.append(curr_states)
logits, value = model(curr_states)
values.append(value.squeeze())
policy = F.softmax(logits, dim=1)
old_m = Categorical(policy)
action = old_m.sample()
actions.append(action)
old_log_policy = old_m.log_prob(action)
old_log_policies.append(old_log_policy)
if torch.cuda.is_available():
[agent_conn.send(("step", act)) for agent_conn, act in zip(envs.agent_conns, action.cpu())]
else:
[agent_conn.send(("step", act)) for agent_conn, act in zip(envs.agent_conns, action)]
state, reward, done, info = zip(*[agent_conn.recv() for agent_conn in envs.agent_conns])
state = torch.from_numpy(np.concatenate(state, 0))
if torch.cuda.is_available():
state = state.cuda()
reward = torch.cuda.FloatTensor(reward)
done = torch.cuda.FloatTensor(done)
else:
reward = torch.FloatTensor(reward)
done = torch.FloatTensor(done)
rewards.append(reward)
dones.append(done)
curr_states = state
_, next_value, = model(curr_states)
next_value = next_value.squeeze()
old_log_policies = torch.cat(old_log_policies).detach()
actions = torch.cat(actions)
values = torch.cat(values).detach()
states = torch.cat(states)
gae = 0
R = []
for value, reward, done in list(zip(values, rewards, dones))[::-1]:
gae = gae * opt.gamma * opt.tau
gae = gae + reward + opt.gamma * next_value.detach() * (1 - done) - value.detach()
next_value = value
R.append(gae + value)
R = R[::-1]
R = torch.cat(R).detach()
advantages = R - values
for i in range(opt.num_epochs):
indice = torch.randperm(opt.num_local_steps * opt.num_processes)
for j in range(opt.batch_size):
batch_indices = indice[
int(j * (opt.num_local_steps * opt.num_processes / opt.batch_size)): int((j + 1) * (
opt.num_local_steps * opt.num_processes / opt.batch_size))]
logits, value = model(states[batch_indices])
new_policy = F.softmax(logits, dim=1)
new_m = Categorical(new_policy)
new_log_policy = new_m.log_prob(actions[batch_indices])
ratio = torch.exp(new_log_policy - old_log_policies[batch_indices])
actor_loss = -torch.mean(torch.min(ratio * advantages[batch_indices],
torch.clamp(ratio, 1.0 - opt.epsilon, 1.0 + opt.epsilon) *
advantages[
batch_indices]))
critic_loss = F.smooth_l1_loss(R[batch_indices], value.squeeze())
entropy_loss = torch.mean(new_m.entropy())
total_loss = actor_loss + critic_loss - opt.beta * entropy_loss
optimizer.zero_grad()
total_loss.backward()
torch.nn.utils.clip_grad_norm_(model.parameters(), 0.5)
optimizer.step()
print("Episode: {}. Total loss: {}".format(curr_episode, total_loss))
if __name__ == "__main__":
opt = get_args()
train(opt)