[docs] dqn report

wbs2788 · Nov 16, 2021 · 34fa757 · 34fa757
1 parent e62fb96
commit 34fa757
Show file tree

Hide file tree

Showing 4 changed files with 7 additions and 13 deletions.
diff --git a/dqn/DQN.py b/dqn/DQN.py
@@ -19,22 +19,21 @@
 
 # Tutorial sample #7: The Maze Decorator
 
+import torch
+
+import game
+import train
+
 try:
     from malmo import MalmoPython
 except:
     import MalmoPython
 
 import json
-import os
-import random
 import sys
 import time
 
 import numpy as np
-import torch
-
-import game
-import train
 
 # from priority_dict import priorityDictionary as PQ
 

diff --git a/dqn/game.py b/dqn/game.py
@@ -7,8 +7,6 @@
 
 import cv2
 import numpy as np
-import random
-import time
 
 class Env():
     def __init__(self):

diff --git a/dqn/train.py b/dqn/train.py
@@ -6,20 +6,17 @@
 @author: godfp
 """
 
-from game import Env
-import time
-import random
 import numpy as np
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
-import tqdm
+
 #参数
 BATCH_SIZE = 32
 LR = 0.01                   # 学习率
 EPSILON = 0.9               # 最优选择动作百分比(有0.9的几率是最大选择，还有0.1是随机选择，增加网络能学到的Q值)
 GAMMA = 0.9                 # 奖励递减参数（衰减作用，如果没有奖励值r=0，则衰减Q值）
-TARGET_REPLACE_ITER = 4   # Q 现实网络的更新频率100次循环更新一次
+TARGET_REPLACE_ITER = 100   # Q 现实网络的更新频率100次循环更新一次
 MEMORY_CAPACITY = 2000      # 记忆库大小
 N_ACTIONS = 4  # 棋子的动作0，1，2，3
 N_STATES = 1

diff --git a/dqn/强化学习第三课作业实验报告.pdf b/dqn/强化学习第三课作业实验报告.pdf