Core layers are the fundamental building blocks of the network. They are responsible for learning relationships between features, shaping data flow, and controlling model complexity.
| Layer | Description | Parameters |
|---|---|---|
| Dense | A fully connected layer that links every input to every output through learnable weights. | Neurons — Number of neurons in this layer. Activation — Function applied to the output. Available options: Linear, ReLU, Sigmoid, Tanh. |
| Dropout | Randomly disables a fraction of neurons during training to reduce overfitting. | Rate — Fraction of neurons that are randomly deactivated during training (for example, 0.2 means 20%). |
| Flatten | Converts multi-dimensional data into a one-dimensional vector to match the input requirements of the next layer. | No parameters required. |
| Reshape | Changes the shape of the input data to ensure compatibility with subsequent layers. | No parameters required. |