initializing weights and biases

example/main.go

@@ -16,6 +16,7 @@ type RequestPayload struct {
 	LearningRate   float64  `json:"learning_rate"`
 	HiddenSize     int      `json:"hidden_size"`
 	ActivationFunc string   `json:"activation"`
+	TestSize       float64  `json:"test_size"`
 }
 
 func main() {
@@ -42,6 +43,7 @@ func main() {
 		LearningRate:   0.01,
 		HiddenSize:     12,
 		ActivationFunc: "tanh",
+		TestSize:       0.3,
 	}
 
 	jsonPayload, err := json.Marshal(payload)

nn/main.go

@@ -7,7 +7,6 @@ import (
 
 	"github.com/go-gota/gota/dataframe"
 	"github.com/gofiber/fiber/v2"
-	"gonum.org/v1/gonum/mat"
 )
 
 type NN struct {
@@ -25,6 +24,10 @@ type NN struct {
 	YTrain dataframe.DataFrame
 	XTest  dataframe.DataFrame
 	YTest  dataframe.DataFrame
+	Wh     [][]float64
+	Bh     []float64
+	Wo     [][]float64
+	Bo     []float64
 }
 
 func NewNN(c *fiber.Ctx) (*NN, error) {
@@ -40,13 +43,15 @@ func NewNN(c *fiber.Ctx) (*NN, error) {
 
 func (nn *NN) Train() {
 	// train test split the data
-	XTrain, XTest, YTrain, YTest := nn.trainTestSplit()
+	_, _, _, _ = nn.trainTestSplit()
 
-	weights, biases := nn.InitWnB()
+	nn.InitWnB()
 
 	// iterate n times where n = nn.Epochs
 	// use backprop algorithm on each iteration
 	// to fit the model to the data
+	for i := 0; i < nn.Epochs; i++ {
+	}
 
 }
 
@@ -54,31 +59,38 @@ func (nn *NN) InitWnB() {
 	// randomly initialize weights and biases to start
 	inputSize := len(nn.Features)
 	hiddenSize := nn.HiddenSize
-	outputSize := 1 // only predicting one thing for now
+	outputSize := 1 // only predicting one thing
+
+	// input hidden layer weights
+	wh := make([][]float64, inputSize)
+	for i := range wh {
+		wh[i] = make([]float64, hiddenSize)
+		for j := range wh[i] {
+			wh[i][j] = rand.Float64() - 0.5
+		}
+	}
 
-	// Initialize weights and biases for the input layer to hidden layer
-	weightsInputHidden := mat.NewDense(inputSize, hiddenSize, nil)
-	weightsInputHidden.Apply(func(_, _ int, v float64) float64 {
-		// Randomly initialize weights with values between -1 and 1
-		return rand.Float64()*2 - 1
-	}, weightsInputHidden)
+	bh := make([]float64, hiddenSize)
+	for i := range bh {
+		bh[i] = rand.Float64() - 0.5
+	}
 
-	biasesHidden := mat.NewVecDense(hiddenSize, nil)
-	biasesHidden.Apply(func(_, _ int, v float64) float64 {
-		// Randomly initialize biases
-		return rand.Float64()
-	}, biasesHidden)
+	// initialize weights and biases for hidden -> output layer
+	wo := make([][]float64, hiddenSize)
+	for i := range wo {
+		wo[i] = make([]float64, outputSize)
+		for j := range wo[i] {
+			wo[i][j] = rand.Float64() - 0.5
+		}
+	}
 
-	// Initialize weights and biases for the hidden layer to output layer
-	weightsHiddenOutput := mat.NewDense(hiddenSize, outputSize, nil)
-	weightsHiddenOutput.Apply(func(_, _ int, v float64) float64 {
-		// Randomly initialize weights with values between -1 and 1
-		return rand.Float64()*2 - 1
-	}, weightsHiddenOutput)
+	bo := make([]float64, outputSize)
+	for i := range bo {
+		bo[i] = rand.Float64() - 0.5
+	}
 
-	biasesOutput := mat.NewVecDense(outputSize, nil)
-	biasesOutput.Apply(func(_, _ int, v float64) float64 {
-		// Randomly initialize biases
-		return rand.Float64()
-	}, biasesOutput)
+	nn.Wh = wh
+	nn.Bh = bh
+	nn.Wo = wo
+	nn.Bo = bo
 }

nn/split.go

@@ -31,10 +31,10 @@ func (nn *NN) trainTestSplit() (dataframe.DataFrame, dataframe.DataFrame, datafr
 	// Create the train DataFrame using the trainIndices
 	train := nn.Df.Subset(trainIndices)
 
-	XTrain = train.Select(nn.Features)
-	YTrain = train.Select(nn.Target)
-	XTest = test.Select(nn.Features)
-	YTest = test.Select(nn.Target)
+	XTrain := train.Select(nn.Features)
+	YTrain := train.Select(nn.Target)
+	XTest := test.Select(nn.Features)
+	YTest := test.Select(nn.Target)
 
 	return XTrain, XTest, YTrain, YTest
 }

server.go

@@ -1,6 +1,8 @@
 package main
 
 import (
+	"fmt"
+
 	"github.com/Jensen-holm/ml-from-scratch/nn"
 	"github.com/gofiber/fiber/v2"
 )
@@ -12,13 +14,16 @@ func main() {
 	// that we will be able to validate.
 	app.Post("/neural-network", func(c *fiber.Ctx) error {
 
-		_, err := nn.NewNN(c)
+		nn, err := nn.NewNN(c)
 		if err != nil {
 			return c.Status(fiber.StatusBadRequest).JSON(fiber.Map{
 				"error": err,
 			})
 		}
 
+		nn.Train()
+		fmt.Println(nn.Wo)
+
 		return c.SendString("No error")
 	})