# Save processor and create model card
tokenizer.save_pretrained(repository_id)
trainer.create_model_card()
trainer.push_to_hub()