Indeed, it sounds like you have what's called fine tuned data (given an input, here's the output), there's loads of info both here on HN about fine tuning and on youtube's huggingface channels
Note if you have sufficient data, look into existing models on huggingface, you may find a smaller, faster and more open (licencing-wise) model that you can fine tune to get the results you want - Llama is hot, but not a catch-all for all tasks (as no model should be)
It's more about quality vs sufficiency - you can have a relatively small but accurate and wide ranging dataset, this is better than an inaccurate huge dataset
Note if you have sufficient data, look into existing models on huggingface, you may find a smaller, faster and more open (licencing-wise) model that you can fine tune to get the results you want - Llama is hot, but not a catch-all for all tasks (as no model should be)
Happy inferring!