之前好几年的编程，遇到各种奇怪的报错，有的解决了有的没有，现在都忘了，我非常痛心。所以从现在（2023.05.10）开始对以后遇到的报错进行记录。
- 1.RuntimeError: Trying to backward through the graph a second time (or directly access saved variables after they have already been freed). Saved intermediate values of the graph are freed when you call .backward() or autograd.grad(). Specify retain_graph=True if you need to backward through the graph a second time or if you need to access saved variables after calling backward.
- 2.File "D:\Projects\Flow\Flow_Forecast\venv\lib\site-packages\psycopg2\__init__.py", line 122, in connect conn = _connect(dsn, connection_factory=connection_factory, **kwasync) psycopg2.OperationalError
- 3.TypeError: can't convert np.ndarray of type numpy.object_. The only supported types are: float64……
- 4.numpy.core._exceptions._ArrayMemoryError: Unable to allocate 3.99 MiB for an array with shape (39, 13410) and data type object
- 5.ufunc 'isnan' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''
1.RuntimeError: Trying to backward through the graph a second time (or directly access saved variables after they have already been freed). Saved intermediate values of the graph are freed when you call .backward() or autograd.grad(). Specify retain_graph=True if you need to backward through the graph a second time or if you need to access saved variables after calling backward.
解决方法：网上较多的是针对一个网络有多个输出分别进行梯度下降的情况，我遇到的则是Softmax和Parameter同时在一条语句中使用时报错，解决方法是去掉Softmax
2.File “D:\Projects\Flow\Flow_Forecast\venv\lib\site-packages\psycopg2_init_.py”, line 122, in connect conn = _connect(dsn, connection_factory=connection_factory, **kwasync) psycopg2.OperationalError
解决方法：这是一个很笼统的报错，以至于我没看出来。后来发现是字母大小写的问题，postgresql使用python新建表单和数据库时大写字母被强制转为小写，但在读取数据库和表的时候却分大小写，所以这边初始化报错了。建议postgresql中的名字都用小写。
3.TypeError: can’t convert np.ndarray of type numpy.object_. The only supported types are: float64……
解决方法：强制类型转换 x.astype(None)，变更为普通数组
4.numpy.core._exceptions._ArrayMemoryError: Unable to allocate 3.99 MiB for an array with shape (39, 13410) and data type object
解决方法：这是读取数据内存溢出的问题。重跑模型或者优化内存配置。
5.ufunc ‘isnan’ not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ‘‘safe’’
解决方法：np.nan(a)中的a的类型有问题，修改为np.nan(a.astype(None))即可。