Check out a few examples that use [~transformers.generation_utils.GenerationMixin.generate] method to produce
text outputs for different tasks:
* Text summarization
* Image captioning
* Audio transcription
Note that the inputs to the generate method depend on the model's modality.