tensorflow keras训练过程中，输出正确率和损失函数值都为0，这个是哪里出问题了

Question

 ### 生成数据集代码：

> 数据集是满足x*x < y的

import random
def line_split():

    with open("dataset/line_split.txt", "w", encoding="utf-8") as f:
        num = 1000
        s = []
        for i in range(num):
            x = random.random()*10
            y = random.random()*20
            if x*x < y:
                if i != (num-1):
                    s.append("{},{},{}\n".format(x, y, 1))
                else:
                    s.append("{},{},{}".format(x, y, 1))
            else:
                if i != (num - 1):
                    s.append("{},{},{}\n".format(x, y, -1))
                else:
                    s.append("{},{},{}".format(x, y, -1))
        f.writelines(s)

### 加载数据代码：

def s2f(num):
    return float(num)

def load_data(file_name):
    with open(file_name, "r", encoding="utf-8") as f:
        lines = f.readlines()
        x, y = [], []
        for line in lines:
            data = list(map(s2f, line.split(",")))
            x.append(data[:-1])
            y.append(data[-1])
        return x, y

### 源代码：

import tensorflow as tf
import load_data

x, y = load_data.load_data("dataset/line_split.txt")
x = tf.cast(x, dtype=tf.float32)
y = tf.cast(y, dtype=tf.int32)

model = tf.keras.models.Sequential([
    tf.keras.layers.Dense(
        2,
        activation="softmax",
        kernel_regularizer=tf.keras.regularizers.l2()  # 正则化
    )]
)

model.compile(
    optimizer=tf.keras.optimizers.SGD(lr=0.5),
    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
    metrics=['binary_accuracy']
)

model.fit(x, y, batch_size=32, epochs=100, validation_split=0.2, validation_freq=20)

model.summary()

### 很奇怪，所有的loss和准确率都为0，这个是为什么？？？

 

Answer 1

请问解决了吗？求帮助