S1quence
S1quence
> 您好,感谢您的工作! > > 我在运行PPO算法时,其中PPOGitHub - model.py - 337行“state = t.from_numpy(state).to(self.device)”报错, TypeError: expected np.ndarray (got tuple) > > 请问是什么问题呢?谢谢您 根本原因->Gym版本问题: 这个 是因为gym版本更新之后,执行 ```python state = env.step(action) ``` 的时候,返回的不是state了,而是一个**tuple**,这就导致了下面代码报错 ```python state...
> 感谢您的回答! 您给的两种方案我都尝试了一下, 方案1报错:AttributeError: 'EntryPoints' object has no attribute 'get' (按照requirements.txt,把环境全换了) 方案2报错:在代码next_state, reward, done, _ = env.step(action.item()),报错ValueError: too many values to unpack (expected 4) > > 再次感谢您的回复! 第一种方案不知道为啥,估计是什么库犯病了,比较麻烦 第二种方案报错是这个原因: 由于GYM库修改了,`env.reset()`和`env.step()`返回的值都加了个`{}`。所以在这里这行代码也应该修改成:...
> May I ask why "'type' object is not subscriptable" appears when running the code?
This is caused by python3.8 or previous versions, I wrote the code using python3.9. You can upgrade the python to 3.9. Here is an example: ```python def func(li:list[str])->list[str]: # this...
> May I ask why "'type' object is not subscriptable" appears when running the code? I have explained the cause of the error. The easiest way is to create a...
> thanks you. Can you give me an example? For example, "def sample(self) -> tuple[np.ndarray, t.ensor, t.ensor, t.ensor, t.ensor, t.ensor, t.ensor]: "What to change to Just delet all the return...
确实搞错了,在这里下载[https://mediatum.ub.tum.de/1554803],我电脑存着一份,需要上传么? 另外还找到了TensorFlow版本https://paperswithcode.com/dataset/sen12ms-cr
嗯好,要等一两天哈,我还得拆成1gb的压缩包,而且百度云速度也很感人,有270gb呢。
> 刚刚下了个压缩包看了下,果真传错了,正在把CR的数据集压成1G压缩包,机械盘比较慢,会尽快上传!!
链接:https://pan.baidu.com/s/11lypfXe24byyk5FM8yZ1ZA 提取码:fxps 终于搞定了一共271GB