Hands-On Deep Learning with Go

Accuracy

As this model takes much longer to converge, we should also add a rudimentary metric to track our accuracy. In order to do this, we must first extract our labels from the data - which we can do as below:

 // get label    yRowT, _ := yVal.Slice(sli{j, j + 1})    yRow := yRowT.Data().([]float64)    var rowLabel int    var yRowHigh float64    for k := 0; k < 10; k++ {        if k == 0 {            rowLabel = 0            yRowHigh = yRow[k]        } else if yRow[k] > yRowHigh {            rowLabel = k            yRowHigh = yRow[k]        }    }

We must then get our prediction from the output data:

yOutput2 := tensor.New(tensor.WithShape(bs, 10), tensor.WithBacking(arrayOutput2)) // get prediction    predRowT, _ := yOutput2.Slice(sli{j, j + 1})    predRow := predRowT.Data().([]float64)    var rowGuess int var predRowHigh float64 ...