- Notifications
You must be signed in to change notification settings - Fork1
Code for ArXiv CS RAG Huggingface Space: Generate scientific paper abstract embeddings using ColBERTv2 and ask questions
NotificationsYou must be signed in to change notification settings
BishmoyPaul/arxiv-CS-RAG
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
This repository hosts the code forArXiv CS RAG, a Huggingface space for searching paper embeddings and querying using large language models (LLMs) of your choice.
ArXivCS.Demo.mp4
- Input a Question: The user inputs a question into the interface.
- Abstract Retrieval: The system uses ColBERTv2 to search ArXiv for the most relevant paper abstracts related to the question.
- Contextual Answer Generation: The retrieved abstracts are then fed into an LLM (Mistral or Gemma-based) to generate a detailed and accurate answer.
- Output: The final answer, along with the relevant abstracts, is displayed to the user.
- Question-Based ArXiv Paper Retrieval: Automatically fetches the most relevant ArXiv paper abstracts by using a question as input.
- ColBERTv2 Retriever: Employs ColBERTv2, a highly efficient retrieval model, to accurately find the most relevant abstracts based on the input question.
- LLM-Powered Answers: Uses advanced LLMs like Mistral or Gemma to generate comprehensive answers grounded in the retrieved paper abstracts.
- Create Embeddings from ArXiv Abstracts | Alternatively, you can check out the codehere too
- Build Huggingface Space