ディープラーニングとGolangの完璧な組み合わせ

package main import ( "fmt" "image" "image/jpeg" "os" "github.com/LdDl/go-deep" ) func main() { // 加载训练好的神经网络模型 model := deep.NewNeural(&deep.Config{ LearningRate: 0.01, ActivationFunc: deep.Sigmoid{}, LossFunc: deep.MeanSquare{}, WeightInitFunc: deep.GaussianInit{}, }) err := model.Load("model.json") if err != nil { fmt.Println("Failed to load model:", err) return } // 加载要分类的图像 file, err := os.Open("test.jpg") if err != nil { fmt.Println("Failed to open image:", err) return } defer file.Close() img, _, err := image.Decode(file) if err != nil { fmt.Println("Failed to decode image:", err) return } // 将图像转换为神经网络可以接受的输入格式 resized := resize(img, 28, 28) inputs := flatten(resized) // 使用训练好的神经网络进行预测 outputs := model.Predict(inputs) // 输出预测结果 fmt.Println("Predicted probabilities:", outputs) } // 缩放图像到指定大小 func resize(img image.Image, width, height int) image.Image { resized := image.NewRGBA(image.Rect(0, 0, width, height)) imgBounds := img.Bounds() scaleX := float64(imgBounds.Max.X) / float64(width) scaleY := float64(imgBounds.Max.Y) / float64(height) for x := 0; x < width; x++ { for y := 0; y < height; y++ { imgX := int(float64(x) * scaleX) imgY := int(float64(y) * scaleY) resized.Set(x, y, img.At(imgX, imgY)) } } return resized } // 将图像像素值展平为一维向量 func flatten(img image.Image) []float64 { bounds := img.Bounds() width, height := bounds.Max.X, bounds.Max.Y var pixels []float64 for y := 0; y < height; y++ { for x := 0; x < width; x++ { r, g, b, _ := img.At(x, y).RGBA() value := (float64(r) + float64(g) + float64(b)) / 3 pixels = append(pixels, value) } } return pixels }