How to generate text: using different decoding methods for language generation with Transformers