Hi! Have you tried implementing dropout layers into your architecture? They're great for preventing overfitting by randomly dropping units from the neural network during training. It might help improve your model's generalization. Also, experimenting with different activation functions like LeakyReLU instead of a standard ReLU might offer some benefits. Keep tweaking, and good luck!
Hi! Have you tried implementing dropout layers into your architecture? They're great for preventing overfitting by randomly dropping units from the neural network during training. It might help improve your model's generalization. Also, experimenting with different activation functions like LeakyReLU instead of a standard ReLU might offer some benefits. Keep tweaking, and good luck!