인공지능 개발 도구에서 자바 프레임워크의 현황은 어떤가요?

import weka.classifiers.Classifier;
import weka.classifiers.Evaluation;
import weka.core.Instances;
import weka.core.converters.ConverterUtils.DataSource;

public class WekaClassifierExample {

    public static void main(String[] args) throws Exception {
        // 加载数据集
        DataSource source = new DataSource("iris.arff");
        Instances data = source.getDataSet();

        // 分离标签
        data.setClassIndex(data.numAttributes() - 1);

        // 训练分类器
        Classifier classifier = new weka.classifiers.bayes.NaiveBayes();
        classifier.buildClassifier(data);

        // 评估分类器
        Evaluation eval = new Evaluation(data);
        eval.evaluateModel(classifier, data);

        // 输出准确率
        System.out.println("准确率：" + eval.pctCorrect());
    }
}

import org.tensorflow.TensorFlow;
import org.tensorflow.keras.layers.Dense;
import org.tensorflow.keras.models.Sequential;
import org.tensorflow.keras.optimizers.Adam;
import org.tensorflow.keras.utils.LayerUtils;
import java.nio.file.Paths;
import java.util.Arrays;

public class TensorFlowExample {

    public static void main(String[] args) {
        // 创建顺序模型
        Sequential model = new Sequential();

        // 添加密集层
        model.add(new Dense(128, activation="relu", inputShape=new int[]{784}));
        model.add(new Dense(10, activation="softmax"));

        // 编译模型
        model.compile(new Adam(0.001f), "categorical_crossentropy", new String[]{"accuracy"});

        // 训练模型
        model.fit(
                Paths.get("mnist_data/train-images-idx3-ubyte").toFile(),
                Paths.get("mnist_data/train-labels-idx1-ubyte").toFile(),
                100,
                32,
                LayerUtils.batchNormalization());

        // 评估模型
        float[] loss = model.evaluate(Paths.get("mnist_data/t10k-images-idx3-ubyte").toFile());
        System.out.println("损失：" + loss[0]);
        System.out.println("准确率：" + loss[1]);
    }
}