CMSI5998

Submissions:

crawl.py
↑ submitted Jul 26

In this file, we crawl Wikepedia until we find 'Jesus'. The data accumulated through this process is pushed to a database.

Proposal.md
↑ submitted Jul 28

In this script, I outline the plan for the final project. Here I choose NewsAPI for finetuning a Llama LLM.

 

Final submissions:

buildDatabase.py
↑ submitted Jul 31

In this submission, I call NewsAPI for news articles relating to key topics in computer science and mathematics. This data is then put into a MongoDB database.

cs_terms.txt
↑ submitted Jul 31

In this text file, a list of key topics in computer science and mathematics are provided.

run_NewsLlama.sh
↑ submitted Jul 31

In this bash file, the pipeline from data accumulation to model evaluation is outlined. This file also sets API key environment variable (redacted for safety).

TerminalView.png
↑ submitted Jul 31

In this image, you can find what the terminal looks like while we build our database.

NewsLlama.py
↑ submitted Jul 31

In this script, we collect the data in the MongoDB database and finetune TinyLlama-- a parameter reduced Llama 2 model.

Everything else ↑ submitted Aug 5