With LSTM appear this kind of circumstance, could you tell me how to improve the

CodePudding user response:

This is typical of fitting, perhaps LSTM is not suitable for your data set,
You also can adjust into the LSTM data length, training methods, or USES two-way LSTM, multilayer LSTM try, multilayer can be added between the LSTM dropout

CodePudding user response:

Ok, I see, thank you ~

CodePudding user response:

Ok, I see, thank you ~