In the loss function 'neuralNDCG', the input 'y_pred' is the prediction from the model, and it's shape is '[batch_size, slate_length]'. What does the 'slate_length' mean? Is each element in 'y_pred' a similarity score or a ranking value? Can you show a example of the loss function inputs?
In the loss function 'neuralNDCG', the input 'y_pred' is the prediction from the model, and it's shape is '[batch_size, slate_length]'. What does the 'slate_length' mean? Is each element in 'y_pred' a similarity score or a ranking value? Can you show a example of the loss function inputs?