Question - How to fix the constant validation accuracy in CNN model training?
Answer -
Constant validation accuracy is a common problem when training any neural network because the network just remembers the sample and results in an overfitting problem. Overfitting of a model means that the neural network model works fantastic on the training sample but the performance of the model sinks in on the validation set. Here are some tips to try to fix the constant validation accuracy in CNN –
- It is always advisable to divide the dataset into training, validation, and test set.
- When working with little data, this problem can be solved by changing the parameters of the neural network by trial and error.
- Increasing the size of the training dataset.
- Use batch normalization.
- Regularization
- Reduce the network complexity