Keras:一键编码的类权重(class_weight) [英] Keras: class weights (class_weight) for one-hot encoding
问题描述
我想在keras model.fit中使用class_weight参数来处理不平衡的训练数据.通过查看一些文档，我了解到我们可以通过这样的字典:
I'd like to use class_weight argument in keras model.fit to handle the imbalanced training data. By looking at some documents, I understood we can pass a dictionary like this:
class_weight = {0 : 1,
1: 1,
2: 5}
(在此示例中，class-2在损失函数中将受到更高的惩罚.)
(In this example, class-2 will get higher penalty in the loss function.)
问题是我的网络输出具有一键编码，即class-0 =(1、0、0)，class-1 =(0、1、0)和class-3 =(0、0， 1).
The problem is that my network's output has one-hot encoding i.e. class-0 = (1, 0, 0), class-1 = (0, 1, 0), and class-3 = (0, 0, 1).
我们如何将class_weight用于一键编码输出?
How can we use the class_weight for one-hot encoded output?
通过查看 Keras中的某些代码，看起来_feed_output_names包含一个输出类列表，但在我的情况下，model.output_names/model._feed_output_names返回['dense_1']
By looking at some codes in Keras, it looks like _feed_output_names contain a list of output classes, but in my case, model.output_names/model._feed_output_names returns ['dense_1']
推荐答案
我想我们可以改用sample_weights.实际上，在Keras内部，class_weights会转换为sample_weights.
I guess we can use sample_weights instead. Inside Keras, actually, class_weights are converted to sample_weights.
sample_weight:与x长度相同的可选数组，包含 应用于每个样本的模型损失的权重.如果是 时间数据，您可以传递带有形状的2D数组(样本， sequence_length)，以对每个时间步长施加不同的权重 每个样本.在这种情况下，您应确保指定 在compile()中为sample_weight_mode ="temporal".
sample_weight: optional array of the same length as x, containing weights to apply to the model's loss for each sample. In the case of temporal data, you can pass a 2D array with shape (samples, sequence_length), to apply a different weight to every timestep of every sample. In this case you should make sure to specify sample_weight_mode="temporal" in compile().
这篇关于Keras:一键编码的类权重(class_weight)的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
