Tensorflow，tf.nn.softmax_cross_entropy_with_logits 和 tf.nn.sparse_softmax_cross_entropy_with_logits 的区别

Question

我已阅读 docs of both functions ，但据我所知，对于函数 tf.nn.softmax_cross_entropy_with_logits(logits, labels, dim=-1, name=None) ，结果是交叉熵损失，其中logits的维度和 labels是相同的。

但是，对于函数 tf.nn.sparse_softmax_cross_entropy_with_logits ，尺寸logits和 labels不一样吗？

你能给出一个更详细的tf.nn.sparse_softmax_cross_entropy_with_logits的例子吗？ ?

Answer 1

不同的是tf.nn.softmax_cross_entropy_with_logits不假设这些类是互斥的:

Measures the probability error in discrete classification tasks in which each class is independent and not mutually exclusive. For instance, one could perform multilabel classification where a picture can contain both an elephant and a dog at the same time.

比较 sparse_* :

Measures the probability error in discrete classification tasks in which the classes are mutually exclusive (each entry is in exactly one class). For example, each CIFAR-10 image is labeled with one and only one label: an image can be a dog or a truck, but not both.

因此，对于稀疏函数， logits 的维度和 labels不一样: labels每个示例包含一个数字，而 logits每个示例的类数，表示概率。