Fine Tuning a model for Prompt Engineering

Hi, I am new to the community. I want to fine-tune a model for Prompt Engineering. I have a few examples of texts and label pairs. The texts explain the symptoms and cause of a disease but do not give the name of the disease, the label is simply the disease name for that text. I have a very small amount of example pairs so I need to create more of these prompts from the few examples that I have. I want to create a model by fine-tuning a pre-trained model for this task. Can anyone guide me in the right direction as to how exactly should I approach this problem? Can I use GPT3 or Bloom or any other model for fine-tuning and prompt engineering?