Beginer: I'm new to machine learning ,fastai and Pytorch. Looking to build a chatbot model that can Predict the answer

I have a csv file with question and answers. I need a model that Can Predict the answer when it matched with the question in csv.
What are the steps that I should follow,

You could start with a pretrained language model. There are many available on HuggingFace, for e.g. GPT-Neo. You could then finetune the model with your questions and answers.



If you are new, I think starting with fastai courses is a good place to build up foundations. The course assumes that you know python.

Hey, Which where can I find the steps to train the model. In my case I have a csv file containing simple question and answers

So the most liked question-answering model on HF at the moment is deepset/roberta-base-squad2.

You can probably just duplicate this notebook and replace the documents in here with your file and format the output as you need.

1 Like