I am new to the field of Neural networks. Currently, I am working on training a CNN model to classify XRAY Images into Normal and Viral Pneumonia.
I am facing an issue of Constant Val accuracy while training the model. However, with each epoch the training accuracy is becoming better and both the losses (loss and Val loss) are decreasing.
How should I deal with this problem of the constant Val Accuracy?
Attaching below: The Screenshot of the Epochs
I agree with Mohammad Deeb. This link is useful: https://stackoverflow.com/questions/52356068/validation-accuracy-constant-in-keras-cnn-for-multiclass-image-classification
I hope you tried with a different validation set. Is the error repeated ?
It seems your model is overfitting. You could try arranging new data or different validation data as mentioned by Shahir Asfahan. If both doesn't work then you could try for opting hyperparameter tuning or a new model altogether.
If you validate your model with different datasets and got this result, it means your model predicts randomly. Keep in mind that all the data you used should be drawn from the same distribution. If the validation data has some extra features and your model hasn't seen these features in training. It can also be the case. As Soumya Snigdha Kundu says, the model is overfitting if the data is correct.
Thanks for adding those valuable points Jian Xi . I'll remember to add them next time!
This is a standard error when training neural networks. Your network remembered the sample. To solve this "overfitting" problem, it is necessary to divide the data set into training, test and validation samples (if there is enough data). If there is little data, then it is necessary to solve this problem by trial and error: change the parameters of the neural network. The criterion for stopping learning: reducing the error of the neural network. All samples must be obtained at random.
I agree with Mohammad Deeb. This link is useful: https://stackoverflow.com/questions/52356068/validation-accuracy-constant-in-keras-cnn-for-multiclass-image-classification
Thank you all for your answers, I tried some of the methods you all suggested and what worked for me was in the link suggested by Mr. @Mohammed Deeb and the answer by Mr. @Dmitry Nazarov . I decreased the number of layers and changed the parameters of the neural networks and it seemed to work.
Dear Palak Nath
This paper may help
Article Novel Feature Selection and Voting Classifier Algorithms for...
- Badges
- Science topic