Choosing the correct model for generate response to a customer's reviews

Hi, my name is Joel, from Argentina. I am working in a project and I am kind of stuck.

My goal is to fine-tunning a “X” model so it generate a response to a customer review. For example:

Given the review: ‘Great storage building with good security.’
It must generate something like this: ‘Thank you for taking the time to leave a review.’

My dataset is composed of “review_by_user” and “response_by_company”

My problem is that I don’t know which kind of model should I use?
I think is GPT2, because it can generate text, but I am not sure about that. Maybe a summarization model can be useful to. So, I am truly new to this topic and I don’t know to identifies the pros and cons of each model that exists. I readed the NLP course that HuggingFace has, I think that the approach for the summary model is maybe similar to my case.

So, I ask for your help and advice. Sorry for my english :stuck_out_tongue: