32 lines
665 B
Python
32 lines
665 B
Python
# %%
|
|
from stable_baselines3 import PPO
|
|
from stable_baselines3.common.env_util import make_vec_env
|
|
from intersim.envs.intersimple import IntersimpleTargetSpeedRandom
|
|
|
|
model_name = "ppo_tspeed_random"
|
|
|
|
# %%
|
|
env = IntersimpleTargetSpeedRandom(target_speed=10)
|
|
|
|
# %%
|
|
model = PPO("MlpPolicy", env, verbose=1)
|
|
model.learn(total_timesteps=250000)
|
|
model.save(model_name)
|
|
|
|
print('Done training.')
|
|
|
|
del model # remove to demonstrate saving and loading
|
|
|
|
# %%
|
|
model = PPO.load(model_name)
|
|
|
|
obs = env.reset()
|
|
while True:
|
|
action, _states = model.predict(obs)
|
|
obs, rewards, done, info = env.step(action)
|
|
env.render(mode='post')
|
|
if done:
|
|
break
|
|
|
|
env.close()
|