This project implements a book recommendation system using Retrieval-Augmented Generation (RAG) techniques. The system retrieves relevant book descriptions from a large dataset and generates recommendations based on user queries using transformers models.
- Text similarity-based document retrieval using
sentence-transformers
. - Text generation using various models from Hugging Face.
- Book recommendations based on user inputs using RAG.
- Utilizes datasets from Hugging Face for corpus generation.
- Retriever:
sentence-transformers/all-MiniLM-L6-v2
- Generators:
microsoft/Phi-3-mini-128k-instruct
,GPT-2
,T5-small
- Python
- Transformers (Hugging Face)
- Sentence-Transformers
- Hugging Face Datasets
- Customize the corpus and queries in the notebook for your own use case.
- Use the retriever to find relevant documents based on user input.
- Generate meaningful recommendations using different text generation models.
The project uses the Goodreads Book Descriptions dataset available on Hugging Face.