from_logits=True

BinaryCrossentropy class

tf.keras.losses.BinaryCrossentropy(
    from_logits=False, label_smoothing=0, reduction="auto", name="binary_crossentropy"
)

Computes the cross-entropy loss between true labels and predicted labels.

Use this cross-entropy loss for binary (0 or 1) classification applications. The loss function requires the following inputs:

• y_true (true label): This is either 0 or 1.
• y_pred (predicted value): This is the model's prediction, i.e, a single floating-point value which either represents a logit, (i.e, value in [-inf, inf] when from_logits=True) or a probability (i.e, value in [0., 1.] when from_logits=False).

搬运说得很清楚了，logit 和 probability的区别：

logit: 输出是

probability:[0,1]

• reduction：类型为tf.keras.losses.Reduction，对loss进行处理，默认是求平均；
• name: op的name

---

CategoricalCrossentropy class

tf.keras.losses.CategoricalCrossentropy(
    from_logits=False,
    label_smoothing=0,
    reduction="auto",
    name="categorical_crossentropy",
)

Computes the crossentropy loss between the labels and predictions.

Use this crossentropy loss function when there are two or more label classes. We expect labels to be provided in a one_hot representation. If you want to provide labels as integers, please use SparseCategoricalCrossentropy loss. There should be # classes floating point values per feature.

In the snippet below, there is # classes floating pointing values per example. The shape of both y_predand y_true are [batch_size, num_classes].

对应二分类，这个categoricalCrossentropy是用于多分类。标签需要是one_hot。如果不是one_hot，可以用SparseCategoricalCrossentropy loss。

---

keras.optimizers.Adam(lr=0.001, beta_1=0.9, beta_2=0.99, epsilon=1e-08, decay=0.0)

• lr：float> = 0.学习率
• beta_1：float，0
• beta_2：float，0
• epsilon：float> = 0,模糊因子。如果None，默认为K.epsilon()。该参数是非常小的数，其为了防止在实现中除以零
• decay：float> = 0,每次更新时学习率下降
   

学习率决定了学习进程的快慢（也可以看作步幅的大小）。如果学习率过大，很可能会越过最优值，反而如果学习率过小，优化的效率可能很低，导致过长的运算时间，Adam大概的思想是开始的学习率设置为一个较大的值，然后根据次数的增多，动态的减小学习率，以实现效率和效果的兼得。