Hi, I am a student, I have written this code for neural network and I’m trying to classify image data using below neural network as covid or viral or normal. But my model, even running on several epochs, it predicts entire images to be either only one of the three classes i.e, if it predicts ,it predicts all images to be covid, or viral or normal. This happens even during the training phase, I think i’m missing something , please help.
the dataset link: COVID-19 Radiography Database | Kaggle
The dataset is imbalanced, but i have also tried taking subset of dataset with equal number of images for each class. Please correct me
Can you share the code and some sample outputs/labels of the model? It might help to get a sense of whether the predictions are not changing during training or something else is happening.