def save(self):
utils.remove_dir(self.path)
utils.make_dir(self.path)
foo = lambda model, params, name: save_model(
model,
os.path.join(self.path, name + MODEL_STATE_EXT),
params
)
self._apply(foo, self.model_names + ['stacker'])
if done:
break
rewards.append(ep_reward)
if ma_rewards:
ma_rewards.append(ma_rewards[-1] * 0.9 + ep_reward * 0.1)
else:
ma_rewards.append(ep_reward)
print(f"回合：{i_ep+1}/{cfg.test_eps}，奖励：{ep_reward:.1f}")
print('完成测试！')
return rewards, ma_rewards
if __name__ == "__main__":
cfg = DQNConfig()
plot_cfg = PlotConfig()
# 训练
env, agent = env_agent_config(cfg, seed=1)
rewards, ma_rewards = train(cfg, env, agent)
make_dir(plot_cfg.result_path, plot_cfg.model_path) # 创建保存结果和模型路径的文件夹
agent.save(path=plot_cfg.model_path) # 保存模型
save_results(rewards, ma_rewards, tag='train',
path=plot_cfg.result_path) # 保存结果
plot_rewards_cn(rewards, ma_rewards, plot_cfg, tag="train") # 画出结果
# 测试
env, agent = env_agent_config(cfg, seed=10)
agent.load(path=plot_cfg.model_path) # 导入模型
rewards, ma_rewards = test(cfg, env, agent)
save_results(rewards, ma_rewards, tag='test',
path=plot_cfg.result_path) # 保存结果
plot_rewards_cn(rewards, ma_rewards, plot_cfg, tag="test") # 画出结果
rewards.append(ep_reward)
if ma_rewards:
ma_rewards.append(0.9 * ma_rewards[-1] + 0.1 * ep_reward)
else:
ma_rewards.append(ep_reward)
print('Complete training！')
return rewards, ma_rewards
if __name__ == "__main__":
cfg = HierarchicalDQNConfig()
# train
env, agent = env_agent_config(cfg, seed=1)
rewards, ma_rewards = train(cfg, env, agent)
make_dir(cfg.result_path, cfg.model_path)
agent.save(path=cfg.model_path)
save_results(rewards, ma_rewards, tag='train', path=cfg.result_path)
plot_rewards(rewards,
ma_rewards,
tag="train",
algo=cfg.algo,
path=cfg.result_path)
# eval
env, agent = env_agent_config(cfg, seed=10)
agent.load(path=cfg.model_path)
rewards, ma_rewards = eval(cfg, env, agent)
save_results(rewards, ma_rewards, tag='eval', path=cfg.result_path)
plot_rewards(rewards,
ma_rewards,
tag="eval",
'/'+curr_time+'/models/' # 保存模型的路径
self.save = True # 是否保存图片
self.device = torch.device(
"cuda" if torch.cuda.is_available() else "cpu") # 检测GPU
def env_agent_config(cfg, seed=1):
env = NormalizedActions(gym.make(cfg.env_name)) # 装饰action噪声
env.seed(seed) # 随机种子
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
agent = DDPG(state_dim, action_dim, cfg)
return env, agent
cfg = DDPGConfig()
plot_cfg = PlotConfig()
# 训练
env, agent = env_agent_config(cfg, seed=1)
rewards, ma_rewards = train(cfg, env, agent)
make_dir(plot_cfg.result_path, plot_cfg.model_path)
agent.save(path=plot_cfg.model_path)
save_results(rewards, ma_rewards, tag='train', path=plot_cfg.result_path)
plot_rewards(rewards, ma_rewards, plot_cfg, tag="train") # 画出结果
# 测试
env, agent = env_agent_config(cfg, seed=10)
agent.load(path=plot_cfg.model_path)
rewards, ma_rewards = test(plot_cfg, env, agent)
save_results(rewards, ma_rewards, tag='test', path=cfg.result_path)
plot_rewards(rewards, ma_rewards, plot_cfg, tag="test") # 画出结果
if done:
break
rewards.append(ep_reward)
if ma_rewards:
ma_rewards.append(ma_rewards[-1] * 0.9 + ep_reward * 0.1)
else:
ma_rewards.append(ep_reward)
print(f"回合：{i_ep+1}/{cfg.test_eps}，奖励：{ep_reward:.1f}")
print('完成测试！')
env.close()
return rewards, ma_rewards
if __name__ == "__main__":
cfg = Config()
# 训练
env, agent = env_agent_config(cfg)
rewards, ma_rewards = train(cfg, env, agent)
make_dir(cfg.result_path, cfg.model_path) # 创建保存结果和模型路径的文件夹
agent.save(path=cfg.model_path) # 保存模型
save_results(rewards, ma_rewards, tag='train',
path=cfg.result_path) # 保存结果
plot_rewards(rewards, ma_rewards, cfg, tag="train") # 画出结果
# 测试
env, agent = env_agent_config(cfg)
agent.load(path=cfg.model_path) # 导入模型
rewards, ma_rewards = test(cfg, env, agent)
save_results(rewards, ma_rewards, tag='test',
path=cfg.result_path) # 保存结果
plot_rewards(rewards, ma_rewards, cfg, tag="test") # 画出结果
curr_path = os.path.dirname(__file__)
parent_path = os.path.dirname(curr_path)
sys.path.append(parent_path) # add current terminal path to sys.path
import gym
import torch
import datetime
from DQN.agent import DQN
from common.plot import plot_rewards
from common.utils import save_results, make_dir, del_empty_dir
SEQUENCE = datetime.datetime.now().strftime(
"%Y%m%d-%H%M%S") # obtain current time
SAVED_MODEL_PATH = curr_path + "/saved_model/" + SEQUENCE + '/' # path to save model
RESULT_PATH = curr_path + "/results/" + SEQUENCE + '/' # path to save rewards
make_dir(curr_path + "/saved_model/", curr_path + "/results/")
del_empty_dir(curr_path + "/saved_model/", curr_path + "/results/")
class DQNConfig:
def __init__(self):
self.env = 'LunarLander-v2'
self.algo = "DQN" # name of algo
self.gamma = 0.95
self.epsilon_start = 1 # e-greedy策略的初始epsilon
self.epsilon_end = 0.01
self.epsilon_decay = 500
self.lr = 0.0001 # learning rate
self.memory_capacity = 1000000 # Replay Memory容量
self.batch_size = 64
self.train_eps = 300 # 训练的episode数目
agent.update()
if i_episode % cfg.target_update == 0:
agent.target_net.load_state_dict(agent.policy_net.state_dict())
print('Episode:{}/{}, Reward:{}'.format(i_episode + 1, cfg.train_eps,
ep_reward))
rewards.append(ep_reward)
# 计算滑动窗口的reward
if ma_rewards:
ma_rewards.append(0.9 * ma_rewards[-1] + 0.1 * ep_reward)
else:
ma_rewards.append(ep_reward)
print('Complete training！')
return rewards, ma_rewards
if __name__ == "__main__":
cfg = DQNConfig()
env = gym.make(cfg.env)
env.seed(1)
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.n
agent = DQN(state_dim, action_dim, cfg)
rewards, ma_rewards = train(cfg, env, agent)
make_dir(cfg.result_path)
agent.save(path=cfg.result_path)
save_results(rewards, ma_rewards, tag='train', path=cfg.result_path)
plot_rewards(rewards,
ma_rewards,
tag="train",
algo=cfg.algo,
path=cfg.result_path)
def get_summary_writer():
name = str(datetime.datetime.now())[:19]
utils.make_dir(PATH['TF_LOGS'])
logs_path = os.path.join(PATH['TF_LOGS'], name)
return SummaryWriter(logs_path)
