-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathtraining_ppo.py
206 lines (180 loc) · 8.17 KB
/
training_ppo.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
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
import numpy as np
import gym_sumo
import gym
from stable_baselines3.common.logger import configure
from stable_baselines3.common.env_util import make_vec_env
from stable_baselines3 import PPO
from wandb.integration.sb3 import WandbCallback
import wandb
import argparse
import os
from pathlib import Path
from gym_sumo.envs import SUMOEnv
from gym_sumo.envs.utils import generateFlowFiles
from matplotlib import pyplot as plt
import warnings
warnings.filterwarnings('ignore')
class SUMOEnvPPO(SUMOEnv):
def __init__(self, reset_callback=None, reward_callback=None, observation_callback=None, info_callback=None, done_callback=None, shared_viewer=True, mode='gui',
edges=..., simulation_end=36000, joint_agents=False, episode_length=20, **kwargs):
super().__init__(reset_callback, reward_callback, observation_callback, info_callback,
done_callback, shared_viewer, mode, edges, simulation_end, joint_agents, **kwargs)
self.action_space = gym.spaces.MultiDiscrete([5, 9, 2])
# observation space
self.observation_space = gym.spaces.Box(
low=0, high=10, shape=(11,), dtype=np.float64)
self.episode_length = episode_length
self._episode_length_counter = 0
def _get_obs(self, agent):
return self.getState(self.edge_agents[0])
def getState(self, edge_agent):
"""
Retrieve the state of the network from sumo.
"""
edge_id = edge_agent.edge_id
normalizeUniqueVehicleCount = 300
laneWidthCar = self.traci.lane.getWidth(f'{edge_id}_2')
laneWidthBike = self.traci.lane.getWidth(f'{edge_id}_1')
laneWidthPed = self.traci.lane.getWidth(f'{edge_id}_0')
nLaneWidthCar = np.interp(laneWidthCar, [0, 12.6], [0, 1])
nLaneWidthBike = np.interp(laneWidthBike, [0, 12.6], [0, 1])
nLaneWidthPed = np.interp(laneWidthPed, [0, 12.6], [0, 1])
# E0 is for agent 0 and 1, #-E0 is for agent 2 and 3, #E1 is for agent 4 and 5, #-E1 is for agent 6 and 7
# E2 is for agent 8 and 9, #-E2 is for agent 10 and 11, #E3 is for agent 12 and 13, #-E3 is for agent 14 and 15
laneVehicleAllowedType = self.traci.lane.getAllowed(f'{edge_id}_0')
if 'bicycle' in laneVehicleAllowedType:
cosharing = 1
else:
cosharing = 0
state = []
state_0 = laneWidthCar
state_1 = laneWidthBike
state_2 = laneWidthPed
state_3 = edge_agent._total_occupancy_car_Lane
state_4 = edge_agent._total_density_car_lane
state_5 = edge_agent._total_occupancy_bike_Lane
state_6 = edge_agent._total_occupancy_ped_Lane
state_7 = float(cosharing) # flag for cosharing on or off
state_8 = float(np.abs(cosharing-1))
state_9 = edge_agent._total_density_bike_lane
state_10 = edge_agent._total_density_ped_lane
state = [state_0, state_1, state_2, state_3, state_4,
state_5, state_6, state_7, state_8, state_9, state_10]
return np.array(state)
def reset(self, *args):
obs = super().reset(*args)
self._episode_length_counter = 0
return list(obs.values())[0]
def step(self, action_n):
obs_n, reward_n, done_n, info_n = super().step(action_n)
self._episode_length_counter += 1
obs_n = list(obs_n.values())[0]
reward_n = reward_n[0]
# if self._episode_length_counter >= 20:
# self._done = True
done_n = self._episode_length_counter >= self.episode_length
return obs_n, reward_n, done_n, info_n
def run(model, config):
model_dir = Path('./models') / config.env_id / config.model_name
curr_run = f'ppo_{env.get_attr("density_threshold")[0]:.2f}_{config.seed}'
# if not model_dir.exists():
# curr_run = 'run1'
# else:
# exst_run_nums = [int(str(folder.name).split('run')[1]) for folder in
# model_dir.iterdir() if
# str(folder.name).startswith('run')]
# if len(exst_run_nums) == 0:
# curr_run = 'run1'
# else:
# curr_run = 'run%i' % (max(exst_run_nums) + 1)
run_dir = model_dir / curr_run
# for ep_i in tqdm(range(0, config.n_episodes)):
# total_reward = 0
# print("Episodes %i-%i of %i" % (ep_i + 1,
# ep_i + 1 + config.n_rollout_threads,
# config.n_episodes))
# # obs = env.reset()
# Train the agent
model.learn(total_timesteps=config.n_episodes*config.episode_length, reset_num_timesteps=True, tb_log_name="PPO", progress_bar=True,
callback=WandbCallback(
model_save_path=run_dir,
verbose=2,))
# step = 0
# for et_i in range(config.episode_length):
# step += 1
# action, _ = model.predict(obs, deterministic=True)
# print("Step {}".format(step + 1))
# print("Action: ", action)
# obs, reward, done, info = env.step(action)
# # model.learn(1)
# print('obs=', obs, 'reward=', reward, 'done=', done)
# if done:
# print("Goal reached!", "reward=", reward)
# break
model.save(run_dir / 'model.pt')
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("--env_id", default="simple", type=str)
parser.add_argument("--model_name", default="simple_model", type=str)
parser.add_argument("--seed",
default=42, type=int,
help="Random seed")
parser.add_argument("--n_rollout_threads", default=1, type=int)
parser.add_argument("--n_training_threads", default=6, type=int)
parser.add_argument("--buffer_length", default=int(1e6), type=int)
parser.add_argument("--n_episodes", default=1500, type=int)
parser.add_argument("--episode_length", default=20, type=int)
parser.add_argument("--steps_per_update", default=10, type=int)
parser.add_argument("--batch_size",
default=1024, type=int,
help="Batch size for model training")
parser.add_argument("--n_exploration_eps", default=24000, type=int)
parser.add_argument("--init_noise_scale", default=0.3, type=float)
parser.add_argument("--final_noise_scale", default=0.0, type=float)
parser.add_argument("--save_interval", default=30, type=int)
parser.add_argument("--hidden_dim", default=64, type=int)
parser.add_argument("--lr", default=0.01, type=float)
parser.add_argument("--tau", default=0.01, type=float)
parser.add_argument("--agent_alg",
default="MADDPG", type=str,
choices=['MADDPG', 'DDPG'])
parser.add_argument("--adversary_alg",
default="MADDPG", type=str,
choices=['MADDPG', 'DDPG'])
parser.add_argument("--discrete_action",
action='store_true')
config = parser.parse_args()
EDGES = ['E0']
generateFlowFiles("Train", edges=EDGES)
joint_agents = False
mode = 'none'
env_kwargs = {'mode': mode,
'edges': EDGES,
'joint_agents': joint_agents}
log_dir = "logs"
if not os.path.exists(log_dir):
os.makedirs(log_dir)
# can be online, offline, or disabled
use_wandb = os.environ.get('WANDB_MODE', 'online')
wandb_run = wandb.init(
project=f"AdaptableLanesRevisionTRC{'PPO_'.lower()}",
tags=["PPO_Final?", "RL"],
mode=use_wandb,
sync_tensorboard=True
)
env_kwargs['episode_length'] = config.episode_length
env = make_vec_env(SUMOEnvPPO, n_envs=4,
seed=config.seed, env_kwargs=env_kwargs)
env.env_method('set_run_mode', 'Train')
# new_logger = configure(log_dir, ["stdout", "csv", "tensorboard"])
# check_env(env, warn=True)
print(env.action_space)
print(env.action_space.sample())
print(env.observation_space)
# env.reset()
model = PPO("MlpPolicy", env, n_steps=20, verbose=1,
tensorboard_log=f'logs/{wandb_run.id}', device='cpu', seed=config.seed)
# model.set_logger(new_logger)
# print(f'logs/{wandb_run.id}')
run(model, config)
wandb_run.finish()