General case question_answering.ipynb about notebooks

[andreabac3]

@Sankalp1233 You can procede in two ways

1. Using a pre-trained QA model: Place your document and the question into the squad dataset format and then feed it into the model to get the answer.
   1.2) To obtain the answer predicted I suggest the reading of this tutorial based on huggingface api.
2. Train your own QA model: In the same way you have to produce a dataset with the triple (context, question, answer).
   2.2) Use this tutorial for training

I am not a huggingface employee I hope I helped you. They are certainly more experienced than me.

[Sankalp1233]

For the tutorial for training I am not getting the code come up. It says Sorry, something went wrong. Reload? and for the other tutorial:https://towardsdatascience.com/simple-and-fast-question-answering-system-using-huggingface-distilbert-single-batch-inference-bcf5a5749571. I have this coming up: How to read this story from Ramsri Goutham — and everything else on Medium.
You’ve read all your free member-only stories. Become a member to get unlimited access and support the voices you want to hear more from.
Upgrade
Read any story. Access everything on Medium across all your devices with no limits or ads.
Reward great writing. A portion of your membership fee will go toward the writers you read most. Can you please try to screenshot the code so I could see it better?

[andreabac3]

Hi @Sankalp1233,
about the tutorial on point (2) is a well-know issue of github.com, so you have to download the jupyter notebook file or clicking here to open it in colab.
Regarding to the towardsdatascience.com try to sign up to medium or opening it via incognito mode in your browser.

[Sankalp1233]

I did open it incognito mode, but I need the part of the article where it states Here is the code to make a single inference with DistilBERT. If you have the code, can you please kindly share it? Does distilbert have the same problem as bert, where it only takes 512 tokens at once?