Summary
The code examples, explanations, and exercises that we covered in this chapter are a quick introduction to training sequence labeling models in Flair, and it should give you the confidence to prepare, train, validate, and use Flair models that solve real-world problems.
Also, you should now be able to have the ability to tell whether model training was a success or whether it should be restarted with different parameters. We learned that there is a large number of parameters that govern sequence labeling model training as well as the training process. There's the learning rate, the number of epochs, the optimizer type, the number of hidden RNN layers, and a long list of other parameters that will likely affect the performance of our model. In the preceding examples, we usually chose the default parameters that Flair happens to use in most of its code examples, but there's no guarantee that the default parameters are the optimal parameters for our problem. In fact...