The problem Learnex solves

“LEARNEX” aims to provide solutions to most of the problems we as students face in this era of digital education.

A)We provide an easy interface where students can either provide an image of their books or notes, links to pages of interests or can directly provide an paragraph and our program will provide an summary of the text (with the option of the length of the summary) which will highly benefit them for reading and last minute revision.The text is extracted using the pytessearct and opencv module.
B)Based on the images or links or paragraphs provided by the user, we provide them with a quiz generated using transformers upon the topic given in the texts provided by them and they can enter their answers.
C)Instead of using normal string matching, we use semantic analysis to match the answers entered by the students with the correct ones so that they do not have to mug up the exact words rather understand the meaning and then answer.

Challenges we ran into

We faced problems while deploying the model streamlit webapp but were successful, however couldnt deploy on heroku due to exceeding size limit.
We first time used transformers for question answer generation.
Due to limited RAM and large size of model, our laptop used to crash which was one of the main hurdles we faced.