Keras model.evaluate() 和 model.predict() 有什么区别? [英] What is the difference between Keras model.evaluate() and model.predict()?

问题描述

我使用 Keras 生物医学图像分割来分割大脑神经元.我使用了 model.evaluate() 它给了我骰子系数:0.916.但是，当我使用 model.predict() 时，然后通过计算 Dice 系数循环预测图像，Dice 系数为 0.82.为什么这两个值不同?

I used Keras biomedical image segmentation to segment brain neurons. I used model.evaluate() it gave me Dice coefficient: 0.916. However, when I used model.predict(), then loop through the predicted images by calculating the Dice coefficient, the Dice coefficient is 0.82. Why are these two values different?

推荐答案

问题在于 Keras 中的每个指标都以下列方式进行评估:

The problem lies in the fact that every metric in Keras is evaluated in a following manner:

1. 对于每个batch，都会评估一个指标值.
2. 损失的当前值(在 k 个批次之后等于计算的 k 个批次的指标的平均值).
3. 最终结果是为所有批次计算的所有损失的平均值.

1. For each batch a metric value is evaluated.
2. A current value of loss (after k batches is equal to a mean value of your metric across computed k batches).
3. The final result is obtained as a mean of all losses computed for all batches.

大多数最流行的指标(如 mse、categorical_crossentropy、mae)等 - 作为每个示例的损失值的平均值- 具有这样的评估最终会得到正确结果的属性.但是在骰子系数的情况下 - 所有批次的平均值不等于在整个数据集上计算的实际值，并且 model.evaluate() 使用这种计算方式 - 这是问题的直接原因.

Most of the most popular metrics (like mse, categorical_crossentropy, mae) etc. - as a mean of loss value of each example - have a property that such evaluation ends up with a proper result. But in case of Dice Coefficient - a mean of its value across all of the batches is not equal to actual value computed on a whole dataset and as model.evaluate() uses such way of computations - this is the direct cause of your problem.

这篇关于Keras model.evaluate() 和 model.predict() 有什么区别?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！