This project was done under the guidance of Prof. Baby Syla for the completion of Masters degree in Computer Applications. This study aims at creating the MCQ's for a given project. As a starting step Im trying to recreate the state of art result. The repository contains the code that I tried for learning mode about this area. The whole project has three main task.
- Generating Question
- Answering those question
- Generating distractors
I focused on the third part, which is to generate the distractors. In in run of studying deeplearning I found it difficult to understand the intution behind the LSTM cells. This project was the end resutl of that study.
Clone this repository If you are using the anaconda, you can directly clone the environment as :
$ conda env create -f environment.yml
$ conda activate tensornlp- Install Spacy
pip install -U spacy - For lemetisation
pip install -U spacy-lookups-data - Download English Data
python -m spacy download en_core_web_sm
- Generate the fill in the blanks
- Generate the distractors for the fill in the blanks
- Generate the wh questions
- Generate the answers for the wh questions
- Generate the distractors for the sequence
The distractor generation part requires the tensorflow = 1.5.0 Upgrade the pip to version> 19
MIT
Free Software, Hell Yeah!
If you feel difficult in understanding LSTM , dont listen to anyone. Go direclty into its equation. Raise an issue if any found.