test_DQN.py

#import Dependencies
import gym
import highway_env
import numpy as np
#!pip install stable-baselines
from stable_baselines3 import HER, SAC, PPO, DQN, A2C, DDPG
#!pip install stable-baselines
import os
os.environ["KMP_DUPLICATE_LIB_OK"]="TRUE"



#Roundabout
env = gym.make('roundabout-v0')
model =DQN('MlpPolicy', env, verbose=1)

for i in range(10):
	done = False
	env.reset()
	while not done:
		env.render()
		action = env.action_space.sample()
		next_state, reward, done, info = env.step(action)
		print(info)
		print(done)
	env.close()

#Create Model
model = DQN('MlpPolicy', env, verbose=1)
model.learn(total_timesteps=10000)




#Save and load model
model.save('roundabout')
del model
model = DQN.load('roundabout')



#Visualize Model
for i in range(10):
	done = False
	obs = env.reset()
	while not done:
		action, _states = model.predict(obs)

		next_state, reward, done, info, _ =env.step(action)
		print(info)
		print(done)
env.close()