-
Notifications
You must be signed in to change notification settings - Fork 1.1k
/
Copy pathtest_dqn.py
163 lines (150 loc) · 6.05 KB
/
test_dqn.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
150
151
152
153
154
155
156
157
158
159
160
161
162
163
import argparse
import os
import gymnasium as gym
import numpy as np
import torch
from torch.utils.tensorboard import SummaryWriter
from tianshou.data import (
Collector,
CollectStats,
PrioritizedVectorReplayBuffer,
ReplayBuffer,
VectorReplayBuffer,
)
from tianshou.env import DummyVectorEnv
from tianshou.policy import DQNPolicy
from tianshou.policy.base import BasePolicy
from tianshou.trainer import OffpolicyTrainer
from tianshou.utils import TensorboardLogger
from tianshou.utils.net.common import Net
from tianshou.utils.space_info import SpaceInfo
def get_args() -> argparse.Namespace:
parser = argparse.ArgumentParser()
parser.add_argument("--task", type=str, default="CartPole-v1")
parser.add_argument("--reward-threshold", type=float, default=None)
parser.add_argument("--seed", type=int, default=1626)
parser.add_argument("--eps-test", type=float, default=0.05)
parser.add_argument("--eps-train", type=float, default=0.1)
parser.add_argument("--buffer-size", type=int, default=20000)
parser.add_argument("--lr", type=float, default=1e-3)
parser.add_argument("--gamma", type=float, default=0.9)
parser.add_argument("--n-step", type=int, default=3)
parser.add_argument("--target-update-freq", type=int, default=320)
parser.add_argument("--epoch", type=int, default=20)
parser.add_argument("--step-per-epoch", type=int, default=10000)
parser.add_argument("--step-per-collect", type=int, default=10)
parser.add_argument("--update-per-step", type=float, default=0.1)
parser.add_argument("--batch-size", type=int, default=64)
parser.add_argument("--hidden-sizes", type=int, nargs="*", default=[128, 128, 128, 128])
parser.add_argument("--training-num", type=int, default=10)
parser.add_argument("--test-num", type=int, default=100)
parser.add_argument("--logdir", type=str, default="log")
parser.add_argument("--render", type=float, default=0.0)
parser.add_argument("--prioritized-replay", action="store_true", default=False)
parser.add_argument("--alpha", type=float, default=0.6)
parser.add_argument("--beta", type=float, default=0.4)
parser.add_argument(
"--device",
type=str,
default="cuda" if torch.cuda.is_available() else "cpu",
)
return parser.parse_known_args()[0]
def test_dqn(args: argparse.Namespace = get_args()) -> None:
env = gym.make(args.task)
assert isinstance(env.action_space, gym.spaces.Discrete)
space_info = SpaceInfo.from_env(env)
args.state_shape = space_info.observation_info.obs_shape
args.action_shape = space_info.action_info.action_shape
if args.reward_threshold is None:
default_reward_threshold = {"CartPole-v1": 195}
args.reward_threshold = default_reward_threshold.get(
args.task,
env.spec.reward_threshold if env.spec else None,
)
# train_envs = gym.make(args.task)
# you can also use tianshou.env.SubprocVectorEnv
train_envs = DummyVectorEnv([lambda: gym.make(args.task) for _ in range(args.training_num)])
# test_envs = gym.make(args.task)
test_envs = DummyVectorEnv([lambda: gym.make(args.task) for _ in range(args.test_num)])
# seed
np.random.seed(args.seed)
torch.manual_seed(args.seed)
train_envs.seed(args.seed)
test_envs.seed(args.seed)
# Q_param = V_param = {"hidden_sizes": [128]}
# model
net = Net(
state_shape=args.state_shape,
action_shape=args.action_shape,
hidden_sizes=args.hidden_sizes,
device=args.device,
# dueling=(Q_param, V_param),
).to(args.device)
optim = torch.optim.Adam(net.parameters(), lr=args.lr)
policy: DQNPolicy = DQNPolicy(
model=net,
optim=optim,
discount_factor=args.gamma,
estimation_step=args.n_step,
target_update_freq=args.target_update_freq,
action_space=env.action_space,
)
# buffer
buf: ReplayBuffer
if args.prioritized_replay:
buf = PrioritizedVectorReplayBuffer(
args.buffer_size,
buffer_num=len(train_envs),
alpha=args.alpha,
beta=args.beta,
)
else:
buf = VectorReplayBuffer(args.buffer_size, buffer_num=len(train_envs))
# collector
train_collector = Collector[CollectStats](policy, train_envs, buf, exploration_noise=True)
test_collector = Collector[CollectStats](policy, test_envs, exploration_noise=True)
# policy.set_eps(1)
train_collector.reset()
train_collector.collect(n_step=args.batch_size * args.training_num)
# log
log_path = os.path.join(args.logdir, args.task, "dqn")
writer = SummaryWriter(log_path)
logger = TensorboardLogger(writer)
def save_best_fn(policy: BasePolicy) -> None:
torch.save(policy.state_dict(), os.path.join(log_path, "policy.pth"))
def stop_fn(mean_rewards: float) -> bool:
return mean_rewards >= args.reward_threshold
def train_fn(epoch: int, env_step: int) -> None:
# eps annnealing, just a demo
if env_step <= 10000:
policy.set_eps(args.eps_train)
elif env_step <= 50000:
eps = args.eps_train - (env_step - 10000) / 40000 * (0.9 * args.eps_train)
policy.set_eps(eps)
else:
policy.set_eps(0.1 * args.eps_train)
def test_fn(epoch: int, env_step: int | None) -> None:
policy.set_eps(args.eps_test)
# trainer
result = OffpolicyTrainer(
policy=policy,
train_collector=train_collector,
test_collector=test_collector,
max_epoch=args.epoch,
step_per_epoch=args.step_per_epoch,
step_per_collect=args.step_per_collect,
episode_per_test=args.test_num,
batch_size=args.batch_size,
update_per_step=args.update_per_step,
train_fn=train_fn,
test_fn=test_fn,
stop_fn=stop_fn,
save_best_fn=save_best_fn,
logger=logger,
).run()
assert stop_fn(result.best_reward)
def test_pdqn(args: argparse.Namespace = get_args()) -> None:
args.prioritized_replay = True
args.gamma = 0.95
args.seed = 1
test_dqn(args)