R Deep Learning Cookbook
上QQ阅读APP看书,第一时间看更新

How it works...

The section covers model performance visualization using TensorBoard. The cross entropy for train and test are recorded in the SCALARS tab, as shown in the following screenshot:

Cross entropy for train and test data

The objective function shows similar behaviors for train and test cost function; thus, the model seems to be stable for the given case with convergence attaining around 1,600 iterations.