boyu-ai/Hands-on-RL

Issues

• 第十四章SAC 算法代码实践中tanh_normal分布的对数概率密度不太对

  #44 opened a year ago by SurprisedCat
  10

• 第九章策略梯度的损失函数

  #79 opened 13 days ago by mgt-lya
  0

• 第八章 `In [7]`代码块，VAnet() 疑似有误

  #54 opened a year ago by Aegis1863
  1

• 用spyder跑PPO代码，kernel自动关闭了

  #67 opened 6 months ago by Shawkncok
  1

• PPO在单摆实验中为什么要对reward=(reward+8)/8的修改呢？

  #70 opened 6 months ago by xxoospring
  2

• https://www.boyuai.com/进不去了

  #78 opened 2 months ago by virtualxiaoman
  0

• 马尔可夫决策过程，MDP转化为MRP时计算的P疑似有误

  #77 opened 2 months ago by zyy777
  0

• DQN和AC算法中的q_targets的loss计算为什么最后要乘（1-done）呢？

  #48 opened a year ago by superbignut
  2

• 关于web教程布局的建议

  #76 opened 3 months ago by dctwan15
  0

• CartPole-v0环境训练reward超过上限值200？

  #72 opened 3 months ago by SHTechBoBo
  1

• 第十三章 DDPG算法 代码实践中的一点疏漏

  #75 opened 3 months ago by xiyanzzz
  0

• 多臂老虎机的代码里面init_prob为什么是1.0?

  #74 opened 3 months ago by mafan1506
  0

• 3.5公式不准确

  #41 opened 3 months ago by administrator418
  0

• 关于环境初始化的一点提示

  #73 opened 4 months ago by Summer907
  0

• 网页版教程 3.3.2 价值函数 推导过程有些模糊

  #71 opened 4 months ago by wangdehua01
  0

• 21章MADDPG代码问题，存在维度不匹配

  #61 opened 9 months ago by CorneliusDeng
  1

• DQN ReplayBuffer

  #68 opened 7 months ago by xxoospring
  1

• 20章的代码问题

  #60 opened 9 months ago by Wayne857
  3

• 第七章DQN代运行报错

  #50 opened a year ago by ShuoZheLi
  3

• 3.5公式错误

  #37 opened a year ago by StevenJokess
  1

• SAC伪代码存在一点小问题

  #69 opened 7 months ago by taojunhui
  0

• 将AC改为off-policy后，每次训练500条左右的经验就会报错，显示action_dist = torch.distributions.Categorical(probs)这行代码的运行结果为tensor([[nan, nan]]

  #66 opened 7 months ago by Chensyfighting
  3

• ValueError: expected sequence of length 3 at dim 2 (got 0)

  #31 opened a year ago by Yang1231
  6

• 7.4 DQN 算法反向传播有没有进行求导？？

  #64 opened 8 months ago by anranyicheng
  1

• 公式错误---https://hrl.boyuai.com/chapter/1/时序差分算法#55-q-learning-算法

  #65 opened 7 months ago by wslgqq277g
  1

• SAC算法——状态价值函数存在问题

  #63 opened 8 months ago by Dilettante258
  0

• 运行环境

  #62 opened 8 months ago by zheng-lv
  1

• 第七章DNQ回报超出200

  #59 opened 9 months ago by KingOfChuXuan
  1

• 已解决

  #58 opened 9 months ago by Thovenfish
  0

• 失业三年人不认可该观点！：UCB的U_t(a)的分母分母中为拉动每根拉杆的次数加上常数 1 ，这确保每个动作**至少被探索一次**

  #57 opened 9 months ago by StevenJokess
  0

• 第7章-DQN算法 训练时报出错误 ValueError: expected sequence of length 4 at dim 2 (got 0)

  #30 opened a year ago by horacehht
  8

• MARL的PPT的第7页和8页参考文献咋相同？

  #56 opened 10 months ago by StevenJokess
  1

• Dueling DQN部分的疑问

  #40 opened a year ago by Ruanzhh
  2

• 第三章 马尔科夫决策过程 3.3.1计算回报的函数有问题

  #55 opened a year ago by Sen1553
  0

• 第8章 拓展阅读公式推导结果有误，补充分部积分过程

  #53 opened a year ago by Aegis1863
  0

• 关于开发环境配置

  #51 opened a year ago by mellody11
  4

• 第9章-策略梯度算法 中的交叉熵损失体现在代码哪里 ？

  #52 opened a year ago by chensisi0730
  0

• 制作了 EPUB 格式

  #49 opened a year ago by wizardforcel
  0

• 蒙特卡罗采样动作和状态 temp变量为什么是累加呢

  #47 opened a year ago by ChengchengDu
  0

• DDPG算法篇笔误

  #46 opened a year ago by Neuerliu
  1

• 第18章cql代码

  #45 opened a year ago by Jaceyxy
  0

• 第十六章 模型预测控制 EnsembleModel类：train方法的问题

  #43 opened a year ago by Yandong23
  1

• 第20章 未定义win?

  #42 opened a year ago by beyondliaaaa
  0

• 网页版本与纸质书的区别？

  #39 opened a year ago by sibangde
  0

• Action Space Limitations in Continuous PPO Algorithm in Chapter 12

  #38 opened a year ago by ASUKaiwenFang
  0

• 14.3公式错误

  #36 opened a year ago by StevenJokess
  0

• 1.3节强化学习的环境中的公式含义不清

  #35 opened a year ago by qixitan
  0

• 2.5公式错误

  #34 opened a year ago by StevenJokess
  1

• 关于使用multiDiscrete acttion spaces的例子

  #33 opened a year ago by jianzuo
  0

• 2.4公式错误

  #32 opened a year ago by StevenJokess
  0