boyu-ai/Hands-on-RL issues and pull requests

#92 - 第10章 actor-critic算法的顺序问题

Issue - State: closed - Opened by ppap36 about 1 month ago - 3 comments

#91 - 请求提供一个requirements.txt

Issue - State: open - Opened by HEHUA2005 about 1 month ago

#90 - 第8章 DDQN代码无法运行

Issue - State: open - Opened by FengYeXuanLv about 1 month ago - 3 comments

#89 - AC cartpole奖励现在破200是正确的吗？

Issue - State: open - Opened by liaojiaxin97 about 2 months ago - 1 comment

#88 - DQN 在 gym 新版本下修改后正常运行，但结果不符合预期，1.8.3 测试是 ok 的

Issue - State: closed - Opened by mango-zx about 2 months ago

#88 - DQN 在 gym 新版本下修改后正常运行，但结果不符合预期，1.8.3 测试是 ok 的

Issue - State: closed - Opened by mango-zx about 2 months ago

#87 - 请教PPO问题

Issue - State: open - Opened by 394262597 3 months ago - 1 comment

#87 - 请教PPO问题

Issue - State: open - Opened by 394262597 3 months ago

#86 - Chapter 7

Issue - State: open - Opened by A1513906286 3 months ago - 1 comment

#86 - Chapter 7

Issue - State: open - Opened by A1513906286 3 months ago - 2 comments

#85 - Chapter 7

Issue - State: open - Opened by A1513906286 3 months ago

#85 - Chapter 7

Issue - State: open - Opened by A1513906286 3 months ago

#84 - 多臂老虎机ε - 贪心算法解释部分有问题

Issue - State: open - Opened by gymdarius 3 months ago

#84 - 多臂老虎机ε - 贪心算法解释部分有问题

Issue - State: open - Opened by gymdarius 3 months ago

#83 - trpo

Issue - State: open - Opened by L-lorish 3 months ago

#83 - trpo

Issue - State: open - Opened by L-lorish 3 months ago

#82 - 策略梯度证明笔误？

Issue - State: open - Opened by lanceyliao 3 months ago - 2 comments

#82 - 策略梯度证明笔误？

Issue - State: open - Opened by lanceyliao 3 months ago - 2 comments

#81 - 第10章Actor-Critic中actor_loss为何加torch.mean？

Issue - State: closed - Opened by lanceyliao 3 months ago

#81 - 第10章Actor-Critic中actor_loss为何加torch.mean？

Issue - State: closed - Opened by lanceyliao 3 months ago

#80 - 3.6. 占用度量，为何逆序计算？

Issue - State: closed - Opened by lanceyliao 4 months ago - 1 comment

#80 - 3.6. 占用度量，为何逆序计算？

Issue - State: closed - Opened by lanceyliao 4 months ago - 1 comment

#79 - 第九章策略梯度的损失函数

Issue - State: open - Opened by mgt-lya 5 months ago - 1 comment

#79 - 第九章策略梯度的损失函数

Issue - State: open - Opened by mgt-lya 5 months ago - 1 comment

#78 - https://www.boyuai.com/进不去了

Issue - State: open - Opened by virtualxiaoman 6 months ago - 1 comment

#78 - https://www.boyuai.com/进不去了

Issue - State: open - Opened by virtualxiaoman 6 months ago - 1 comment

#77 - 马尔可夫决策过程，MDP转化为MRP时计算的P疑似有误

Issue - State: open - Opened by zyy777 6 months ago - 1 comment

#77 - 马尔可夫决策过程，MDP转化为MRP时计算的P疑似有误

Issue - State: open - Opened by zyy777 6 months ago - 1 comment

#76 - 关于web教程布局的建议

Issue - State: open - Opened by dctwan15 7 months ago

#76 - 关于web教程布局的建议

Issue - State: open - Opened by dctwan15 7 months ago

#75 - 第十三章 DDPG算法代码实践中的一点疏漏

Issue - State: open - Opened by xiyanzzz 8 months ago

#75 - 第十三章 DDPG算法代码实践中的一点疏漏

Issue - State: open - Opened by xiyanzzz 8 months ago

#74 - 多臂老虎机的代码里面init_prob为什么是1.0?

Issue - State: open - Opened by mafan1506 8 months ago

#74 - 多臂老虎机的代码里面init_prob为什么是1.0?

Issue - State: open - Opened by mafan1506 8 months ago

#73 - 关于环境初始化的一点提示

Issue - State: open - Opened by Summer907 8 months ago

#73 - 关于环境初始化的一点提示

Issue - State: open - Opened by Summer907 8 months ago

#72 - CartPole-v0环境训练reward超过上限值200？

Issue - State: closed - Opened by SHTechBoBo 8 months ago - 1 comment

#72 - CartPole-v0环境训练reward超过上限值200？

Issue - State: closed - Opened by SHTechBoBo 8 months ago - 1 comment

#71 - 网页版教程 3.3.2 价值函数推导过程有些模糊

Issue - State: open - Opened by wangdehua01 8 months ago

#71 - 网页版教程 3.3.2 价值函数推导过程有些模糊

Issue - State: open - Opened by wangdehua01 8 months ago

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

Issue - State: closed - Opened by xxoospring 10 months ago - 2 comments

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

Issue - State: closed - Opened by xxoospring 10 months ago - 2 comments

#69 - SAC伪代码存在一点小问题

Issue - State: open - Opened by taojunhui 11 months ago

#69 - SAC伪代码存在一点小问题

Issue - State: open - Opened by taojunhui 11 months ago

#68 - DQN ReplayBuffer

Issue - State: open - Opened by xxoospring 11 months ago - 1 comment

#68 - DQN ReplayBuffer

Issue - State: open - Opened by xxoospring 11 months ago - 1 comment

#67 - 用spyder跑PPO代码，kernel自动关闭了

Issue - State: closed - Opened by Shawkncok 11 months ago - 1 comment

#67 - 用spyder跑PPO代码，kernel自动关闭了

Issue - State: closed - Opened by Shawkncok 11 months ago - 1 comment

#66 - 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]]

Issue - State: open - Opened by Chensyfighting 11 months ago - 3 comments

#66 - 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]]

Issue - State: open - Opened by Chensyfighting 11 months ago - 3 comments

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

Issue - State: closed - Opened by wslgqq277g 12 months ago - 1 comment

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

Issue - State: closed - Opened by wslgqq277g 12 months ago - 1 comment

#64 - 7.4 DQN 算法反向传播有没有进行求导？？

Issue - State: open - Opened by anranyicheng about 1 year ago - 1 comment

#64 - 7.4 DQN 算法反向传播有没有进行求导？？

Issue - State: open - Opened by anranyicheng about 1 year ago - 1 comment

#63 - SAC算法——状态价值函数存在问题

Issue - State: open - Opened by Dilettante258 about 1 year ago

#63 - SAC算法——状态价值函数存在问题

Issue - State: open - Opened by Dilettante258 about 1 year ago

#62 - 运行环境

Issue - State: open - Opened by zheng-lv about 1 year ago - 1 comment

#62 - 运行环境

Issue - State: open - Opened by zheng-lv about 1 year ago - 1 comment

#61 - 21章MADDPG代码问题，存在维度不匹配

Issue - State: open - Opened by CorneliusDeng about 1 year ago - 2 comments

#61 - 21章MADDPG代码问题，存在维度不匹配

Issue - State: open - Opened by CorneliusDeng about 1 year ago - 2 comments

#60 - 20章的代码问题

Issue - State: open - Opened by Wayne857 about 1 year ago - 3 comments

#60 - 20章的代码问题

Issue - State: open - Opened by Wayne857 about 1 year ago - 3 comments

#59 - 第七章DNQ回报超出200

Issue - State: closed - Opened by KingOfChuXuan about 1 year ago - 1 comment

#59 - 第七章DNQ回报超出200

Issue - State: closed - Opened by KingOfChuXuan about 1 year ago - 1 comment

#58 - 已解决

Issue - State: closed - Opened by Thovenfish about 1 year ago

#58 - 已解决

Issue - State: closed - Opened by Thovenfish about 1 year ago

#57 - 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

Issue - State: open - Opened by StevenJokess about 1 year ago

#57 - 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

Issue - State: open - Opened by StevenJokess about 1 year ago

#56 - MARL的PPT的第7页和8页参考文献咋相同？

Issue - State: open - Opened by StevenJokess about 1 year ago - 1 comment

#56 - MARL的PPT的第7页和8页参考文献咋相同？

Issue - State: open - Opened by StevenJokess about 1 year ago - 1 comment

#55 - 第三章马尔科夫决策过程 3.3.1计算回报的函数有问题

Issue - State: open - Opened by Sen1553 over 1 year ago

#55 - 第三章马尔科夫决策过程 3.3.1计算回报的函数有问题

Issue - State: open - Opened by Sen1553 over 1 year ago

#54 - 第八章 `In [7]`代码块，VAnet() 疑似有误

Issue - State: open - Opened by Aegis1863 over 1 year ago - 1 comment

#54 - 第八章 `In [7]`代码块，VAnet() 疑似有误

Issue - State: open - Opened by Aegis1863 over 1 year ago - 1 comment

#53 - 第8章拓展阅读公式推导结果有误，补充分部积分过程

Issue - State: open - Opened by Aegis1863 over 1 year ago

#53 - 第8章拓展阅读公式推导结果有误，补充分部积分过程

Issue - State: open - Opened by Aegis1863 over 1 year ago

#52 - 第9章-策略梯度算法中的交叉熵损失体现在代码哪里？

Issue - State: open - Opened by chensisi0730 over 1 year ago

#52 - 第9章-策略梯度算法中的交叉熵损失体现在代码哪里？

Issue - State: open - Opened by chensisi0730 over 1 year ago

#51 - 关于开发环境配置

Issue - State: open - Opened by mellody11 over 1 year ago - 4 comments

#51 - 关于开发环境配置

Issue - State: open - Opened by mellody11 over 1 year ago - 4 comments

#50 - 第七章DQN代运行报错

Issue - State: open - Opened by ShuoZheLi over 1 year ago - 3 comments

#50 - 第七章DQN代运行报错

Issue - State: open - Opened by ShuoZheLi over 1 year ago - 3 comments

#49 - 制作了 EPUB 格式

Issue - State: open - Opened by wizardforcel over 1 year ago

#49 - 制作了 EPUB 格式

Issue - State: open - Opened by wizardforcel over 1 year ago

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

Issue - State: open - Opened by superbignut over 1 year ago - 2 comments

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

Issue - State: open - Opened by superbignut over 1 year ago - 2 comments

#47 - 蒙特卡罗采样动作和状态 temp变量为什么是累加呢

Issue - State: open - Opened by ChengchengDu over 1 year ago - 1 comment

#47 - 蒙特卡罗采样动作和状态 temp变量为什么是累加呢

Issue - State: open - Opened by ChengchengDu over 1 year ago - 1 comment

#46 - DDPG算法篇笔误

Issue - State: closed - Opened by Neuerliu over 1 year ago - 1 comment

#46 - DDPG算法篇笔误

Issue - State: closed - Opened by Neuerliu over 1 year ago - 1 comment

#45 - 第18章cql代码

Issue - State: open - Opened by Jaceyxy over 1 year ago

#45 - 第18章cql代码

Issue - State: open - Opened by Jaceyxy over 1 year ago

#44 - 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

Issue - State: open - Opened by SurprisedCat over 1 year ago - 11 comments

#44 - 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

Issue - State: open - Opened by SurprisedCat over 1 year ago - 11 comments

#43 - 第十六章模型预测控制 EnsembleModel类：train方法的问题

Issue - State: open - Opened by Yandong23 over 1 year ago - 1 comment

#43 - 第十六章模型预测控制 EnsembleModel类：train方法的问题

Issue - State: open - Opened by Yandong23 over 1 year ago - 1 comment

#42 - 第20章未定义win?

Issue - State: closed - Opened by beyondliaaaa over 1 year ago

#42 - 第20章未定义win?

Issue - State: closed - Opened by beyondliaaaa over 1 year ago

#41 - 3.5公式不准确

Issue - State: closed - Opened by administrator418 over 1 year ago

#41 - 3.5公式不准确

Issue - State: closed - Opened by administrator418 over 1 year ago

GitHub / boyu-ai/Hands-on-RL issues and pull requests