update

codes/A2C/task0.py

@@ -123,14 +123,15 @@ def train(cfg,envs):
         loss.backward()
         optimizer.step()
     print('Finish training！')
-    return test_rewards, test_ma_rewards
+    return {'rewards':test_rewards,'ma_rewards':test_ma_rewards}
 if __name__ == "__main__":
     cfg = get_args()
     envs = [make_envs(cfg.env_name) for i in range(cfg.n_envs)]
     envs = SubprocVecEnv(envs) 
     # training
-    rewards,ma_rewards = train(cfg,envs)
+    res_dic = train(cfg,envs)
     make_dir(cfg.result_path,cfg.model_path)
     save_args(cfg)
-    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
+    save_results(res_dic, tag='train',
+                 path=cfg.result_path)
+    plot_rewards(res_dic['rewards'], res_dic['ma_rewards'], cfg, tag="train") # 画出结果

codes/DDPG/task0.py

@@ -5,7 +5,7 @@
 @Email: johnjim0816@gmail.com
 @Date: 2020-06-11 20:58:21
 @LastEditor: John
-LastEditTime: 2022-07-21 00:05:41
+LastEditTime: 2022-07-21 21:51:34
 @Discription: 
 @Environment: python 3.7.7
 '''
@@ -86,7 +86,7 @@ def train(cfg, env, agent):
         else:
             ma_rewards.append(ep_reward)
     print('Finish training!')
-    return rewards, ma_rewards
+    return {'rewards':rewards,'ma_rewards':ma_rewards}
 
 def test(cfg, env, agent):
     print('Start testing')
@@ -111,21 +111,23 @@ def test(cfg, env, agent):
             ma_rewards.append(ep_reward)
         print(f"Epside:{i_ep+1}/{cfg.test_eps}, Reward:{ep_reward:.1f}")
     print('Finish testing!')
-    return rewards, ma_rewards
+    return {'rewards':rewards,'ma_rewards':ma_rewards}
 if __name__ == "__main__":
     cfg = get_args()
     # training
     env,agent = env_agent_config(cfg,seed=1)
-    rewards, ma_rewards = train(cfg, env, agent)
+    res_dic = train(cfg, env, agent)
     make_dir(cfg.result_path, cfg.model_path)
     save_args(cfg)
     agent.save(path=cfg.model_path)
-    save_results(rewards, ma_rewards, tag='train', path=cfg.result_path)
-    plot_rewards(rewards, ma_rewards, cfg, tag="train")  
+    save_results(res_dic, tag='train',
+                 path=cfg.result_path)  
+    plot_rewards(res_dic['rewards'], res_dic['ma_rewards'], cfg, tag="train")  
     # testing
     env,agent = env_agent_config(cfg,seed=10)
     agent.load(path=cfg.model_path)
-    rewards,ma_rewards = test(cfg,env,agent)
-    save_results(rewards,ma_rewards,tag = 'test',path = cfg.result_path)
-    plot_rewards(rewards, ma_rewards, cfg, tag="test")  
+    res_dic = test(cfg,env,agent)
+    save_results(res_dic, tag='test',
+                 path=cfg.result_path)  
+    plot_rewards(res_dic['rewards'], res_dic['ma_rewards'], cfg, tag="test")   
 

codes/DQN/task0.py

@@ -10,7 +10,7 @@ import torch
 import datetime
 import numpy as np
 import argparse
-from common.utils import save_results_1, make_dir
+from common.utils import save_results, make_dir
 from common.utils import plot_rewards,save_args
 from dqn import DQN
 
@@ -95,8 +95,8 @@ def train(cfg, env, agent):
 
 
 def test(cfg, env, agent):
-    print('开始测试!')
-    print(f'环境：{cfg.env_name}, 算法：{cfg.algo_name}, 设备：{cfg.device}')
+    print('Start testing!')
+    print(f'Env:{cfg.env_name}, A{cfg.algo_name}, 设备：{cfg.device}')
     ############# 由于测试不需要使用epsilon-greedy策略，所以相应的值设置为0 ###############
     cfg.epsilon_start = 0.0  # e-greedy策略中初始epsilon
     cfg.epsilon_end = 0.0  # e-greedy策略中的终止epsilon
@@ -123,7 +123,7 @@ def test(cfg, env, agent):
         else:
             ma_rewards.append(ep_reward)
         print(f'Episode：{i_ep+1}/{cfg.test_eps}, Reward:{ep_reward:.2f}, Step:{ep_step:.2f}')
-    print('完成测试！')
+    print('Finish testing')
     env.close()
     return {'rewards':rewards,'ma_rewards':ma_rewards,'steps':steps}
 
@@ -133,16 +133,16 @@ if __name__ == "__main__":
     # 训练
     env, agent = env_agent_config(cfg)
     res_dic = 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
-    save_args(cfg)
-    agent.save(path=cfg.model_path)  # 保存模型
-    save_results_1(res_dic, tag='train',
-                 path=cfg.result_path)  # 保存结果
-    plot_rewards(res_dic['rewards'], res_dic['ma_rewards'], cfg, tag="train")  # 画出结果
+    make_dir(cfg.result_path, cfg.model_path)  
+    save_args(cfg) # save parameters
+    agent.save(path=cfg.model_path)  # save model
+    save_results(res_dic, tag='train',
+                 path=cfg.result_path)  
+    plot_rewards(res_dic['rewards'], res_dic['ma_rewards'], cfg, tag="train")  
     # 测试
     env, agent = env_agent_config(cfg)
     agent.load(path=cfg.model_path)  # 导入模型
     res_dic = test(cfg, env, agent)
-    save_results_1(res_dic, tag='test',
+    save_results(res_dic, tag='test',
                  path=cfg.result_path)  # 保存结果
     plot_rewards(res_dic['rewards'], res_dic['ma_rewards'],cfg, tag="test")  # 画出结果

codes/DoubleDQN/outputs/CartPole-v0/20220721-000842/results/params.json

@@ -1 +0,0 @@
-{"algo_name": "DoubleDQN", "env_name": "CartPole-v0", "train_eps": 200, "test_eps": 20, "gamma": 0.99, "epsilon_start": 0.95, "epsilon_end": 0.01, "epsilon_decay": 500, "lr": 0.0001, "memory_capacity": 100000, "batch_size": 64, "target_update": 2, "hidden_dim": 256, "device": "cuda", "result_path": "C:\\Users\\24438\\Desktop\\rl-tutorials\\codes\\DoubleDQN/outputs/CartPole-v0/20220721-000842/results/", "model_path": "C:\\Users\\24438\\Desktop\\rl-tutorials\\codes\\DoubleDQN/outputs/CartPole-v0/20220721-000842/models/", "save_fig": true}

codes/DoubleDQN/outputs/CartPole-v0/20220721-215416/results/params.json

@@ -0,0 +1,19 @@
+{
+    "algo_name": "DoubleDQN",
+    "env_name": "CartPole-v0",
+    "train_eps": 200,
+    "test_eps": 20,
+    "gamma": 0.99,
+    "epsilon_start": 0.95,
+    "epsilon_end": 0.01,
+    "epsilon_decay": 500,
+    "lr": 0.0001,
+    "memory_capacity": 100000,
+    "batch_size": 64,
+    "target_update": 2,
+    "hidden_dim": 256,
+    "device": "cuda",
+    "result_path": "C:\\Users\\24438\\Desktop\\rl-tutorials\\codes\\DoubleDQN/outputs/CartPole-v0/20220721-215416/results/",
+    "model_path": "C:\\Users\\24438\\Desktop\\rl-tutorials\\codes\\DoubleDQN/outputs/CartPole-v0/20220721-215416/models/",
+    "save_fig": true
+}

codes/DoubleDQN/task0.py

@@ -5,7 +5,7 @@ Author: JiangJi
 Email: johnjim0816@gmail.com
 Date: 2021-11-07 18:10:37
 LastEditor: JiangJi
-LastEditTime: 2022-07-21 00:08:38
+LastEditTime: 2022-07-21 21:52:31
 Discription: 
 '''
 import sys,os
@@ -86,7 +86,7 @@ def train(cfg,env,agent):
         else:
             ma_rewards.append(ep_reward)   
     print('Finish training!')
-    return rewards,ma_rewards
+    return {'rewards':rewards,'ma_rewards':ma_rewards}
 
 def test(cfg,env,agent):
     print('Start testing')
@@ -115,22 +115,24 @@ def test(cfg,env,agent):
             ma_rewards.append(ep_reward)
         print(f"Epside:{i_ep+1}/{cfg.test_eps}, Reward:{ep_reward:.1f}")
     print('Finish testing!')
-    return rewards,ma_rewards  
+    return {'rewards':rewards,'ma_rewards':ma_rewards}
 
 if __name__ == "__main__":
     cfg = get_args()
     print(cfg.device)
     # training
     env,agent = env_agent_config(cfg,seed=1)
-    rewards, ma_rewards = train(cfg, env, agent)
+    res_dic = train(cfg, env, agent)
     make_dir(cfg.result_path, cfg.model_path)
     save_args(cfg)
     agent.save(path=cfg.model_path)
-    save_results(rewards, ma_rewards, tag='train', path=cfg.result_path)
-    plot_rewards(rewards, ma_rewards, cfg, tag="train")  
+    save_results(res_dic, tag='train',
+                 path=cfg.result_path)  
+    plot_rewards(res_dic['rewards'], res_dic['ma_rewards'], cfg, tag="train")  
     # testing
     env,agent = env_agent_config(cfg,seed=10)
     agent.load(path=cfg.model_path)
-    rewards,ma_rewards = test(cfg,env,agent)
-    save_results(rewards,ma_rewards,tag = 'test',path = cfg.result_path)
-    plot_rewards(rewards, ma_rewards, cfg, tag="test")  
+    res_dic = test(cfg,env,agent)
+    save_results(res_dic, tag='test',
+                 path=cfg.result_path)  
+    plot_rewards(res_dic['rewards'], res_dic['ma_rewards'], cfg, tag="test") 

codes/PolicyGradient/task0.py

@@ -5,56 +5,47 @@ Author: John
 Email: johnjim0816@gmail.com
 Date: 2020-11-22 23:21:53
 LastEditor: John
-LastEditTime: 2022-02-10 06:13:21
+LastEditTime: 2022-07-21 21:44:00
 Discription: 
 Environment: 
 '''
-import sys
-import os
-curr_path = os.path.dirname(os.path.abspath(__file__))  # 当前文件所在绝对路径
-parent_path = os.path.dirname(curr_path)  # 父路径
-sys.path.append(parent_path)  # 添加路径到系统路径
+import sys,os
+curr_path = os.path.dirname(os.path.abspath(__file__))  # current path
+parent_path = os.path.dirname(curr_path)  # parent path
+sys.path.append(parent_path)  # add to system path
 
 import gym
 import torch
 import datetime
+import argparse
 from itertools import count
 
 from pg import PolicyGradient
 from common.utils import save_results, make_dir
 from common.utils import plot_rewards
 
-curr_time = datetime.datetime.now().strftime("%Y%m%d-%H%M%S")  # 获取当前时间
 
-class Config:
-    '''超参数
-    '''
-
-    def __init__(self):
-        ################################## 环境超参数 ###################################
-        self.algo_name = "PolicyGradient"  # 算法名称
-        self.env_name = 'CartPole-v0' # 环境名称
-        self.device = torch.device(
-            "cuda" if torch.cuda.is_available() else "cpu")  # 检测GPUgjgjlkhfsf风刀霜的撒发十
-        self.seed = 10 # 随机种子，置0则不设置随机种子
-        self.train_eps = 300 # 训练的回合数
-        self.test_eps = 30 # 测试的回合数
-        ################################################################################
-        
-        ################################## 算法超参数 ###################################
-        self.batch_size = 8 # mini-batch SGD中的批量大小
-        self.lr = 0.01 # 学习率
-        self.gamma = 0.99 # 强化学习中的折扣因子
-        self.hidden_dim = 36 # 网络隐藏层
-        ################################################################################
-        
-        ################################# 保存结果相关参数 ################################
-        self.result_path = curr_path + "/outputs/" + self.env_name + \
-            '/' + curr_time + '/results/'  # 保存结果的路径
-        self.model_path = curr_path + "/outputs/" + self.env_name + \
-            '/' + curr_time + '/models/'  # 保存模型的路径
-        self.save = True # 是否保存图片
-        ################################################################################
+def get_args():
+    """ Hyperparameters
+    """
+    curr_time = datetime.datetime.now().strftime("%Y%m%d-%H%M%S")  # Obtain current time
+    parser = argparse.ArgumentParser(description="hyperparameters")      
+    parser.add_argument('--algo_name',default='PolicyGradient',type=str,help="name of algorithm")
+    parser.add_argument('--env_name',default='CartPole-v0',type=str,help="name of environment")
+    parser.add_argument('--train_eps',default=300,type=int,help="episodes of training")
+    parser.add_argument('--test_eps',default=20,type=int,help="episodes of testing")
+    parser.add_argument('--gamma',default=0.99,type=float,help="discounted factor")
+    parser.add_argument('--lr',default=0.01,type=float,help="learning rate")
+    parser.add_argument('--batch_size',default=8,type=int)
+    parser.add_argument('--hidden_dim',default=36,type=int)
+    parser.add_argument('--device',default='cpu',type=str,help="cpu or cuda") 
+    parser.add_argument('--result_path',default=curr_path + "/outputs/" + parser.parse_args().env_name + \
+            '/' + curr_time + '/results/' )
+    parser.add_argument('--model_path',default=curr_path + "/outputs/" + parser.parse_args().env_name + \
+            '/' + curr_time + '/models/' ) # path to save models
+    parser.add_argument('--save_fig',default=True,type=bool,help="if save figure or not")           
+    args = parser.parse_args()                          
+    return args
 
 
 def env_agent_config(cfg,seed=1):
@@ -65,9 +56,9 @@ def env_agent_config(cfg,seed=1):
     return env,agent
 
 def train(cfg,env,agent):
-    print('开始训练!')
-    print(f'环境：{cfg.env_name}, 算法：{cfg.algo_name}, 设备：{cfg.device}')
-    state_pool = [] # 存放每batch_size个episode的state序列
+    print('Start training!')
+    print(f'Env:{cfg.env_name}, Algorithm:{cfg.algo_name}, Device:{cfg.device}')
+    state_pool = [] # temp states pool per several episodes
     action_pool = []
     reward_pool = [] 
     rewards = []
@@ -86,11 +77,11 @@ def train(cfg,env,agent):
             reward_pool.append(reward)
             state = next_state
             if done:
-                print('回合：{}/{}, 奖励：{}'.format(i_ep + 1, cfg.train_eps, ep_reward))
+                print(f'Episode：{i_ep+1}/{cfg.train_eps}, Reward:{ep_reward:.2f}')
                 break
         if i_ep > 0 and i_ep % cfg.batch_size == 0:
             agent.update(reward_pool,state_pool,action_pool)
-            state_pool = [] # 每个episode的state
+            state_pool = [] 
             action_pool = []
             reward_pool = []
         rewards.append(ep_reward)
@@ -99,8 +90,8 @@ def train(cfg,env,agent):
                 0.9*ma_rewards[-1]+0.1*ep_reward)
         else:
             ma_rewards.append(ep_reward)
-    print('完成训练！')
-    env.close()
+    print('Finish training!')
+    env.close() # close environment
     return rewards, ma_rewards
             
 

codes/common/utils.py

@@ -5,7 +5,7 @@ Author: John
 Email: johnjim0816@gmail.com
 Date: 2021-03-12 16:02:24
 LastEditor: John
-LastEditTime: 2022-07-20 23:53:34
+LastEditTime: 2022-07-21 21:45:33
 Discription: 
 Environment: 
 '''
@@ -69,19 +69,19 @@ def plot_losses(losses, algo="DQN", save=True, path='./'):
         plt.savefig(path+"losses_curve")
     plt.show()
 
-def save_results_1(dic, tag='train', path='./results'):
+def save_results(dic, tag='train', path='./results'):
     ''' 保存奖励
     '''
     for key,value in dic.items():
         np.save(path+'{}_{}.npy'.format(tag,key),value)
     print('Results saved！')
     
-def save_results(rewards, ma_rewards, tag='train', path='./results'):
-    ''' 保存奖励
-    '''
-    np.save(path+'{}_rewards.npy'.format(tag), rewards)
-    np.save(path+'{}_ma_rewards.npy'.format(tag), ma_rewards)
-    print('Result saved!')
+# def save_results(rewards, ma_rewards, tag='train', path='./results'):
+#     ''' 保存奖励
+#     '''
+#     np.save(path+'{}_rewards.npy'.format(tag), rewards)
+#     np.save(path+'{}_ma_rewards.npy'.format(tag), ma_rewards)
+#     print('Result saved!')
 
 
 def make_dir(*paths):