I am building DNNRegression model with 2 inputs (i.e. X data) & 2 outputs (i.e. Y data) column data. When I create my input function with these inputs & outputs it gives me following error.
Error: ValueError: Wrong number of items passed 2, placement implies 1.
I know that error is because of 2 columns of "output" values which I am assigning to "Y" through y_train. So is DNNRegression of TensorFlow limited to predict only one column output? Is there any way to predict all output columns?