GIS-PuppetMaster

- [ ] I have marked all applicable categories: + [ ] exception-raising bug + [ ] RL algorithm bug + [ ] documentation request (i.e. "X is missing from...

Bug of PPO

4

ratio = tf.exp(pi.log_prob(action) - old_pi.log_prob(action)) surr = ratio * adv ... loss = -tf.reduce_mean( tf.minimum(surr, tf.clip_by_value(ratio, 1. - self.epsilon, 1. + self.epsilon) * adv) ) should use ratio in tf.minimum...

Request to implement ACKTR algorithm in this link： https://github.com/openai/baselines/tree/f2729693253c0ef4d4086231d36e0a4307ec1cb3/baselines/acktr

rt，也可以在问题右下角三个点的菜单中加入[加入浮窗]的功能

没看懂为什么push里面要把ba的类型转成np.int64

1

`v_wrap(np.array(ba), dtype=np.int64) if ba[0].dtype == np.int64 else v_wrap(np.vstack(ba)),` 一般连续动作不应该用float64吗 而且判断条件是dtype==np.int64时转为np.int64，是不是写错了？

Does RLzoo support Dict gym env state?

2

My customized gym env has a dict type obs_space, even I also customized ActorNetwork and CriticNetwork, I found out RLzoo's source code seems only support single input and can not...

请求新增规则 战争游戏：红龙

1

**需要增加游戏\程序名字** Wargame:RedDragon 请使用[自动抓IP工具](https://github.com/oooldtoy/SSTAP_ip_crawl_tool)来帮助我们编写规则，请详细写明游戏细节，如哪个服，哪个版本。 **该游戏\程序的运行环境** 如 PC **该游戏\程序哪部分需要代理** 如 整个游戏\程序都需要 **特殊情况** 如果还有特殊想说明的请在这里描述.

严重卡顿

1

不知道是不是动画太慢了，电脑是new surface pro 58256,系统RS2，有全套VS和AS,应该不是缺少运行库什么的导致的

rt

Show more custom information of tensor and gradient

1

1. add support to show multiple lines attribute value 2. allowed to customize which attr of the variable can be displayed