Weird test accuracy for VGG-like training

Deep Learning
22

The hunch was that: X_normalized and y_one_hot are being created by appending classes 1 to 5. Hence it might be that the top 80% of data consists of classes 1 to 4, and the bottom 20% has class 5.

Since the validation_split was set to 20%, and since Keras extracts the last n% of your dataset for validation, our training set might contain classes 1 to 4, and validation set with just class 5.

Which means during training, the model would not have seen class 5 at all, hence the poor validation accuracy.

To check this out, solution was to shuffle the contents of X_normalized and y_one_hot so that classes are uniformly distributed.

–

23

Hi everyone
The problem was solved by shuffling data manually before training model as @anandsaha suggested. Here is the first result after fixing this problem:

solved zero accuracy

Now I am going to improve my work. Please let me know if you have any suggestion to improve validation accuracy

Thanks