update rainbowdqn
This commit is contained in:
johnjim0816
@@ -1,5 +1,7 @@
|
||||
import sys
|
||||
import os
|
||||
import torch.nn as nn
|
||||
import torch.nn.functional as F
|
||||
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
|
||||
@@ -8,26 +10,42 @@ import gym
|
||||
import torch
|
||||
import datetime
|
||||
import numpy as np
|
||||
from common.utils import save_results, make_dir
|
||||
from common.utils import save_results_1, make_dir
|
||||
from common.utils import plot_rewards
|
||||
from DQN.dqn import DQN
|
||||
from dqn import DQN
|
||||
|
||||
curr_time = datetime.datetime.now().strftime("%Y%m%d-%H%M%S") # 获取当前时间
|
||||
|
||||
class MLP(nn.Module):
|
||||
def __init__(self, n_states,n_actions,hidden_dim=128):
|
||||
""" 初始化q网络，为全连接网络
|
||||
n_states: 输入的特征数即环境的状态维度
|
||||
n_actions: 输出的动作维度
|
||||
"""
|
||||
super(MLP, self).__init__()
|
||||
self.fc1 = nn.Linear(n_states, hidden_dim) # 输入层
|
||||
self.fc2 = nn.Linear(hidden_dim,hidden_dim) # 隐藏层
|
||||
self.fc3 = nn.Linear(hidden_dim, n_actions) # 输出层
|
||||
|
||||
def forward(self, x):
|
||||
# 各层对应的激活函数
|
||||
x = F.relu(self.fc1(x))
|
||||
x = F.relu(self.fc2(x))
|
||||
return self.fc3(x)
|
||||
|
||||
class Config:
|
||||
'''超参数
|
||||
'''
|
||||
|
||||
def __init__(self):
|
||||
################################## 环境超参数 ###################################
|
||||
self.algo_name = 'DQN' # 算法名称
|
||||
self.env_name = 'CartPole-v0' # 环境名称
|
||||
############################### hyperparameters ################################
|
||||
self.algo_name = 'DQN' # algorithm name
|
||||
self.env_name = 'CartPole-v0' # environment name
|
||||
self.device = torch.device(
|
||||
"cuda" if torch.cuda.is_available() else "cpu") # 检测GPUgjgjlkhfsf风刀霜的撒发十
|
||||
"cuda" if torch.cuda.is_available() else "cpu") # check GPU
|
||||
self.seed = 10 # 随机种子，置0则不设置随机种子
|
||||
self.train_eps = 200 # 训练的回合数
|
||||
self.test_eps = 30 # 测试的回合数
|
||||
self.test_eps = 20 # 测试的回合数
|
||||
################################################################################
|
||||
|
||||
################################## 算法超参数 ###################################
|
||||
@@ -41,8 +59,8 @@ class Config:
|
||||
self.target_update = 4 # 目标网络的更新频率
|
||||
self.hidden_dim = 256 # 网络隐藏层
|
||||
################################################################################
|
||||
|
||||
################################# 保存结果相关参数 ##############################
|
||||
|
||||
################################# 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 + \
|
||||
@@ -55,9 +73,11 @@ def env_agent_config(cfg):
|
||||
''' 创建环境和智能体
|
||||
'''
|
||||
env = gym.make(cfg.env_name) # 创建环境
|
||||
state_dim = env.observation_space.shape[0] # 状态维度
|
||||
action_dim = env.action_space.n # 动作维度
|
||||
agent = DQN(state_dim, action_dim, cfg) # 创建智能体
|
||||
n_states = env.observation_space.shape[0] # 状态维度
|
||||
n_actions = env.action_space.n # 动作维度
|
||||
print(f"n states: {n_states}, n actions: {n_actions}")
|
||||
model = MLP(n_states,n_actions)
|
||||
agent = DQN(n_actions, model, cfg) # 创建智能体
|
||||
if cfg.seed !=0: # 设置随机种子
|
||||
torch.manual_seed(cfg.seed)
|
||||
env.seed(cfg.seed)
|
||||
@@ -72,10 +92,13 @@ def train(cfg, env, agent):
|
||||
print(f'环境：{cfg.env_name}, 算法：{cfg.algo_name}, 设备：{cfg.device}')
|
||||
rewards = [] # 记录所有回合的奖励
|
||||
ma_rewards = [] # 记录所有回合的滑动平均奖励
|
||||
steps = []
|
||||
for i_ep in range(cfg.train_eps):
|
||||
ep_reward = 0 # 记录一回合内的奖励
|
||||
ep_step = 0
|
||||
state = env.reset() # 重置环境，返回初始状态
|
||||
while True:
|
||||
ep_step += 1
|
||||
action = agent.choose_action(state) # 选择动作
|
||||
next_state, reward, done, _ = env.step(action) # 更新环境，返回transition
|
||||
agent.memory.push(state, action, reward,
|
||||
@@ -87,16 +110,18 @@ def train(cfg, env, agent):
|
||||
break
|
||||
if (i_ep + 1) % cfg.target_update == 0: # 智能体目标网络更新
|
||||
agent.target_net.load_state_dict(agent.policy_net.state_dict())
|
||||
steps.append(ep_step)
|
||||
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)
|
||||
if (i_ep + 1) % 10 == 0:
|
||||
print('回合：{}/{}, 奖励：{}'.format(i_ep + 1, cfg.train_eps, ep_reward))
|
||||
print('完成训练！')
|
||||
if (i_ep + 1) % 1 == 0:
|
||||
print(f'Episode：{i_ep+1}/{cfg.test_eps}, Reward:{ep_reward:.2f}, Step:{ep_step:.2f} Epislon:{agent.epsilon(agent.frame_idx):.3f}')
|
||||
print('Finish training!')
|
||||
env.close()
|
||||
return rewards, ma_rewards
|
||||
res_dic = {'rewards':rewards,'ma_rewards':ma_rewards,'steps':steps}
|
||||
return res_dic
|
||||
|
||||
|
||||
def test(cfg, env, agent):
|
||||
@@ -108,41 +133,45 @@ def test(cfg, env, agent):
|
||||
################################################################################
|
||||
rewards = [] # 记录所有回合的奖励
|
||||
ma_rewards = [] # 记录所有回合的滑动平均奖励
|
||||
steps = []
|
||||
for i_ep in range(cfg.test_eps):
|
||||
ep_reward = 0 # 记录一回合内的奖励
|
||||
ep_step = 0
|
||||
state = env.reset() # 重置环境，返回初始状态
|
||||
while True:
|
||||
ep_step+=1
|
||||
action = agent.choose_action(state) # 选择动作
|
||||
next_state, reward, done, _ = env.step(action) # 更新环境，返回transition
|
||||
state = next_state # 更新下一个状态
|
||||
ep_reward += reward # 累加奖励
|
||||
if done:
|
||||
break
|
||||
steps.append(ep_step)
|
||||
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(f'Episode：{i_ep+1}/{cfg.train_eps}, Reward:{ep_reward:.2f}, Step:{ep_step:.2f}')
|
||||
print('完成测试！')
|
||||
env.close()
|
||||
return rewards, ma_rewards
|
||||
return {'rewards':rewards,'ma_rewards':ma_rewards,'steps':steps}
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
cfg = Config()
|
||||
# 训练
|
||||
env, agent = env_agent_config(cfg)
|
||||
rewards, ma_rewards = train(cfg, env, agent)
|
||||
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
|
||||
agent.save(path=cfg.model_path) # 保存模型
|
||||
save_results(rewards, ma_rewards, tag='train',
|
||||
save_results_1(res_dic, tag='train',
|
||||
path=cfg.result_path) # 保存结果
|
||||
plot_rewards(rewards, ma_rewards, cfg, tag="train") # 画出结果
|
||||
plot_rewards(res_dic['rewards'], res_dic['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',
|
||||
res_dic = test(cfg, env, agent)
|
||||
save_results_1(res_dic, tag='test',
|
||||
path=cfg.result_path) # 保存结果
|
||||
plot_rewards(rewards, ma_rewards, cfg, tag="test") # 画出结果
|
||||
plot_rewards(res_dic['rewards'], res_dic['ma_rewards'],cfg, tag="test") # 画出结果
|
||||
|
||||
Reference in New Issue
Block a user