How QNode is learning and calculating gradient?

I want to know how QNode is learning and calculating gradient in the below link
Is this QNode learning calculating gradient on the principle of variational classifier?

Hey @mustafaec,

The QNode object is transformed to a Keras layer by using qml.qnn.KerasLayer(qnode, weight_shapes, output_dim=n_qubits) it creates a TensorFlow compatible tape and then the model is trained with the Keras SGD optimizer. The QNode comes with a specific diff_method=best which will be used to evaluate the gradient in the background for the given device. You can find more information about QNode in the documentation here.

Feel free to ask more questions if needed!