Question

我正在使用Tensorflow Hub进行二进制文本分类。

fit

当我使用fit_generator命令拟合模型时，可以将历元设置为所需的最大值。当我使用# Trains as expected model.fit(x_train, y_train, epochs=100) # Errors during training with # 'Your dataset iterator ran out of data; interrupting training. # Make sure that your iterator can generate at least `steps_per_epoch * epochs` # batches (in this case, 600 batches).' model.fit_generator(my_iterator(x_train, y_train), epochs=100, steps_per_epoch=len(x_train))方法时，似乎每个数据点只能使用一次。

<?php 
$_SESSION = array();
session_destroy();
header("Location: index.php?p=main");

alert("You have been successfully logged out!");

如何正确设置迭代器以使用生成器进行训练？

Answer 1

“预计生成器将无限期地循环其数据”。因此需要将我的迭代器定义为：

    def my_iterator(x, y):
    while True:
        for _x, _y in zip(x, y):
            yield np.array([_x]), np.array([_y])

Tensorflow Fit_generator-如何正确使用纪元？

1 个答案: