-
Streamlit Deployment
-
Hugging Face Deployment
-
Methodology (1)
-
Methodology (2)
-
Evaluation Metrics
Inspiration
When you need to know quickly what choices of modules you have that fit your interests, it will be a time-consuming task to actually read long, detailed texts of module descriptions one by one. That's the problem that I'm trying to solve.
What it does
It's to simplify the long, detailed texts of college modules into just a few quick and easily-grasped tags/keywords that represent the whole original text.
How we built it
It used KeyBERT algorithm as keyphrase extractor algorithm to determine which words or phrases that are significantly more important and relevant. I tried 3 different approaches, then decide which is the best using some evaluation metrics such as Precision, MRR, and MAP.
Challenges we ran into
Tried to fine-tune the BERT transformer, but was unable to do it right now.
Accomplishments that we're proud of
It works quite well :D
What we learned
Goood things
What's next for College Module Tag Generator
Automation
Built With
- hugging-face
- python
- streamlit
Log in or sign up for Devpost to join the conversation.